The umbrella subject is written on your assignment sheet. The work is getting from there down to a question that takes a side.
The topic is not free in this course. The student who tries to build the chain backwards gets turned away in week 4.
In ENG 102 the research paper is not free topic. One broad subject area is set for the section, and everyone pulls their own question out of THAT area. The reason is practical: so your instructor can run tutorials around the same literature, so the presentations feed each other, and so the source pool is shared.
An umbrella like Happiness cannot be written about on its own: it is not arguable, it is not measurable and the number of sources is endless. The whole job is getting from that umbrella down to a question with a side to it, one you can close with five or six days of library work.
Your own section's umbrella subject is written on the assignment sheet: the "Broad Subject area" line in the RESEARCH PROJECT task specification document. Read it there, do not guess. We have no evidence about other sections' umbrellas, which is why there is no list here. Carry your question outside the umbrella and you lose points on the Task Requirements item.
Everything below is a worked example. The same line stands in both of the documents we have, from two separate terms: "Broad Subject area: Happiness". So the topic is not that term's whim, it has been used as the course's settled subject. The reading texts of the essay stage are under the same umbrella.
The reading texts of the essay stage are under the same umbrella too: newspaper and research center pieces on the relationship between happiness and income, plus a 1968 philosophy article arguing about the definition of happiness. So the umbrella sets not only the research paper but the term's readings.
The method works the same under any umbrella, Happiness is only the example. The seven steps below, the 23 real questions and the narrowing table work in the same order under your own umbrella too.
The question texts shared in the same class on the same day. Who wrote them is not recorded. They are here because this is the most concrete answer to "how do I write a question": completely dissimilar questions can come out of the same umbrella.
Next to each step is how that step actually played out in the file we measured.
| Step | What actually happened |
|---|---|
| Verify the umbrella off the assignment sheet | Broad Subject area: Happiness |
| Multiply your interest by the umbrella | The interest was artificial intelligence, the umbrella was happiness, the intersection: is AI raising happiness. |
| Put the question into the "to what extent" pattern | The first version, written in the class group, was "to what extent ai leads into happiness"; in the outline it was cleaned up to "To What Extent Artificial Intelligence Affects Happiness?". |
| Test the source count first, then commit to the question | 23 PDFs in total were collected at the intersection of AI and happiness, and 10 of them were eventually cited. Two thirds got cut, and that is normal. |
| Use the annotated bibliography as a filter | 6 sources were annotated in the annotated bibliography but only 5 of them got an APA citation. Of those 6 sources, only 2 made it into the research draft's works cited. |
| Split the thesis into three reasons and turn the reasons into outline headings | The thesis: reforms/regulations plus health plus education. In the outline, III, IV and V are those three items; II is the counterargument and VI the conclusion. |
| Get the draft out early, then repair at sentence level | Between two drafts a day apart, changes were made in 11 places, and all but two of them were transition and citation fixes. |
What the right hand column has in common: in every one of them you know what to look for. Go to the library with the question "is it harmful" and you will not know what to search.
| Broad topic | First version | Why it does not work | An answerable version |
|---|---|---|---|
| Social media | Is social media harmful to adolescents? | "Harm" is undefined, there is no sample, the answer cannot be measured | Why do large scale studies measuring the relationship between adolescent screen use and mental health give contradictory results? |
| AI and work | Will artificial intelligence take our jobs? | A prediction about the future cannot be closed with evidence | Why do forecasts that measure automation risk at occupation level reach a different conclusion from studies that measure it at task level? |
| Climate policy | Does a carbon tax work? | Which country, which year, which criterion: all left open | In countries that apply carbon pricing, how much of the emission decline do studies attribute to the pricing, and why do these estimates diverge? |
| The city and traffic | Is pedestrianization a good idea? | A value judgment, with no criterion | What do studies measuring shop revenue on pedestrianized streets actually measure, and how does that choice change the result? |
| Scientific publishing | Can science be trusted? | The scope is endless | After the large replication studies in psychology, was there a measurable change in journals' publication practices? |
If the question is too broad, tighten one of three things, not all of them at once.
The 100 entries below do NOT replace the umbrella. First read your own section's Broad Subject area line, then look at the theme closest to it. If your umbrella is "happiness" you can narrow from the health, education, work and city themes; if it is "work and technology", from the technology and automation themes. Go into week 4 with an entry that does not sit under the umbrella and you get turned away, and two weeks are gone.
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| Algorithmic decisions and the right to object | When an algorithm says 'no' to you, can you object? The focus is not the accuracy of the decision but its auditability. | Obermeyer's study shows bias in a real health algorithm with numbers; Dressel and Farid do the same job on criminal risk prediction and find that a simple model is as good as the commercial software. So you do not need to say 'bad', you argue about the measurement. | O'Neil 2016 · Obermeyer 2019 · Dressel 2018 · Mehrabi 2021 |
| How model bias is measured | Saying 'the algorithm is biased' is easy. But under which definition? Where do definitions of fairness collide mathematically? | Caliskan makes human biases in word embeddings measurable, Dwork formalizes the definition of fairness, and Model Cards brings a proposed solution. Three different kinds of source: measurement, theory, practice. | Caliskan 2017 · Dwork 2012 · Mitchell 2019 · Benjamin 2019 |
| Generative AI: where the risk is | Not 'models produce false information' but: which risk grows as the scale grows, and which one shrinks? | Stochastic Parrots sorts the risks into categories and is open access. Taddeo and Floridi defend the other side, the benefit, with an institutional frame. Mitchell teaches you to measure the hype. One of the few topics where you can find an example from your own course. | Bender 2021 · Taddeo 2018 · Mitchell 2019 · Russell 2019 |
| 'I have nothing to hide' | Taking the most common objection to the privacy argument seriously and refuting it from the inside. | Schneier goes after the objection directly; Gorwa gives you the state and platform regulation side. Easy to build a thesis on, hard to defend, and that is a good kind of hard. | Schneier 2015 · Zuboff 2019 · Gorwa 2024 |
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| Why the automation forecast did not hold | ⚠ THIS TOPIC IS WORKED THROUGH IN FULL ON THIS PAGE: the thesis, the claim and support table and a sample paragraph are written above, so it cannot be handed in as it stands. Take the move, then turn the topic into your own with one of the three narrowing levers. In 2013 it was estimated that 47% of total US employment was at risk of automation within twenty years. What happened ten years later, and which assumption did the forecast get wrong? | The forecast, its critique and the later data are all in your hands: a three legged, genuinely falsifiable question. The occupation level versus task level distinction becomes the backbone of the paper. Citation trap: the Oxford Martin working paper is dated 2013 and the peer reviewed version came out in 2017; state which one you are citing. | Frey 2013 · Frey 2017 · Autor 2015 · Acemoglu 2017 |
| Does technology destroy work or rearrange it | ⚠ THIS TOPIC IS WORKED THROUGH IN FULL ON THIS PAGE: the thesis, the claim and support table and a sample paragraph are written above, so it cannot be handed in as it stands. Take the move, then turn the topic into your own with one of the three narrowing levers. The same technology has both a displacement effect and a new task creation effect. What determines the net result? | Acemoglu and Restrepo model and measure these two effects separately; Power and Progress is the book version of the same argument. Taking both your numerical source and your theoretical frame from the same pair of authors is a weakness, and saying so yourself in the paper gains you ground. | Acemoglu 2019 · Acemoglu 2017 · Acemoglu 2023 · Autor 2015 |
| The local effect of trade: the 'China shock' | The gains from global trade show up in the average, the costs in particular regions. How is that asymmetry measured? | Autor, Dorn and Hanson measure it at the level of the local labor market, a classic study that takes a macro argument down to micro data. Standing gives you the sociological side of the same phenomenon. | Autor 2013 · Standing 2011 · Acemoglu 2023 |
| Does the minimum wage reduce employment | The textbook says it does, the data is mixed. Where is the difference: in the theory or in the measurement? | Cengiz and colleagues' method (looking at the bottom end of the job count distribution) is a clean example of how measurement design determines the result. If you can explain the method, the paper goes well above average. | Cengiz 2019 · Spiegelhalter 2019 · Wheelan 2013 |
| The meaning of work | Is people finding their own job meaningless a subjective state, or a measurable pattern? | Graeber's own data is survey based and has been criticized, so the methods argument can sit at the center of the paper. Terkel gives historical depth for comparison, Standing a structural frame. | Graeber 2018 · Terkel 1974 · Standing 2011 |
| Unseen labor | When domestic and care work go unmeasured, how does the economic picture change? | Hochschild takes on the mechanism, Criado Perez the data gap itself. A rare topic that teaches you to use the fact of 'no data' as evidence. | Hochschild 1989 · Perez 2019 · Graeber 2018 |
| Is the digital economy really different | What is the measurable part of the claim that 'digital changed everything'? | Goldfarb and Tucker reduce the difference to five economic properties, turning the slogan into testable items. Zervas gives a concrete case, Benkler the production side. | Goldfarb 2019 · Zervas 2017 · Benkler 2006 |
| The platform economy and labor | Is a platform an intermediary or an employer? Which criterion settles that? | Srnicek gives the theoretical frame, Brynjolfsson-McAfee and Ford the technology and employment side. All three sit in different places on the optimist and pessimist axis, so the counterargument is ready made. | Srnicek 2017 · Brynjolfsson 2014 · Ford 2015 |
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| The attention economy | How is the cost of a free service measured? Turning the slogan 'you are the product' into a measurable claim. | Staying at slogan level is easy, which is what makes this distinctive. Goldfarb and Tucker count which properties of the digital economy are genuinely different in the language of economics, and that is the way to tie the slogan to theory. | Wu 2016 · Zuboff 2019 · Goldfarb 2019 · Lanier 2018 |
| Adolescent screen use and mental health | ⚠ THIS TOPIC IS WORKED THROUGH IN FULL ON THIS PAGE: the thesis, the claim and support table and a sample paragraph are written above, so it cannot be handed in as it stands. Take the move, then turn the topic into your own with one of the three narrowing levers. Not who is right: why do two sides looking at the same data reach different conclusions? | The sides are named, the evidence is open, the argument is not closed. Careful: Odgers's piece is a book review, not a primary study, and mixing that up is the most common mistake. | Haidt 2024 · Orben 2019 · Odgers 2024 · Przybylski 2017 · Twenge 2017 |
| Is the filter bubble real | Separating whether polarization comes from the algorithm or from people's own choices. | Pariser's 2011 claim weakened in later empirical work; Del Vicario and Allcott-Gentzkow measure it with data. Tracking how a claim erodes over time makes a good backbone for a paper. | Pariser 2011 · Vicario 2016 · Allcott 2017 |
| Why falsehood spreads fast | Why is correcting a lie more expensive than spreading it? The focus is not content but speed of circulation. | Vosoughi's Science study measures the spread rate directly; the Council of Europe report hands you the concept map (mis-/dis-/mal-information) ready made. You can start work without getting into a definitional argument. | Vosoughi 2018 · Wardle 2017 · Bergstrom 2020 |
| Is the scale of fake news exaggerated | The public says 'fake news is everywhere'. What is its share of total news consumption? | Allen and colleagues measure exactly that share and find it small; Allcott-Gentzkow point the same way. A thesis that runs against the common view and can be backed with evidence, a move your instructor has not seen a hundred times. | Allen 2020 · Allcott 2017 · Vosoughi 2018 |
| Content moderation on platforms | Can a platform be 'neutral'? The focus is not the censorship argument but the architecture of decision making. | Gillespie gives you how the decisions get made, Gorwa the politics of regulation. Together they turn 'who decides' into an institutional question. | Gillespie 2018 · Gorwa 2024 · Tufekci 2017 |
| Networked protest: power or fragility | Could social media be making mobilization easier while weakening organization? | Tufekci carries field observation from both Turkey and other contexts and builds exactly this paradox; Benkler gives the production side, Herman-Chomsky the classic media model. | Tufekci 2017 · Benkler 2006 · Herman 1988 |
| Entertainment and public discourse | Does the argument Postman built for television hold for today's short video? | Carrying an old thesis into a new medium is both original and controlled. You have to show which part of Postman's argument carries over, and that sharpens the thesis. | Postman 1985 · McLuhan 1964 · Herman 1988 |
| The language of the internet | Is written language changing on the internet a decay, or a new system? | McCulloch examines this linguistically, Pinker argues about the nature of language, Deutscher about language's effect on thought. A clean topic that turns 'young people are ruining the language' into a measurable question. | McCulloch 2019 · Pinker 1994 · Deutscher 2010 |
| Is distraction individual or structural | Is the trouble with focus a matter of personal discipline, or a designed outcome? | Hari builds the structural claim and has been criticized, which is good material for a methods argument. Wu and Zuboff give the economic mechanism. Your own screen time is not data; do not use it as evidence in the paper. | Hari 2022 · Wu 2016 · Zuboff 2019 |
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| Same data, different method: how much does the result move | ⚠ THIS TOPIC IS WORKED THROUGH IN FULL ON THIS PAGE: the thesis, the claim and support table and a sample paragraph are written above, so it cannot be handed in as it stands. Take the move, then turn the topic into your own with one of the three narrowing levers. How much can a researcher's reasonable choices of method change the result? The focus is not the findings but their fragility. | Orben and Przybylski run hundreds of thousands of reasonable analyses on the same data set and show the range of the result; Vuorre tests change over time; Head gives the scale of p-hacking. The most concrete topic for a methods argument. | Orben 2019 · Vuorre 2021 · Head 2015 · Spiegelhalter 2019 |
| The replication crisis | When a published finding cannot be repeated, whose failure is it, the researcher's or the incentive system's? | Ioannidis's 2005 article is open access and short; the 2015 replication study answers it with numbers; Baker shows the field's own admission through a survey. The source chain is clean. | Ioannidis 2005 · Open Science Collaboration 2015 · Baker 2016 · Munafò 2017 |
| What 'significant' means: the p value argument | When did a threshold becoming an instrument of scientific decision turn into a problem? | Amrhein's Nature comment gives the state of the argument, Head measures how widespread p-hacking is, Spiegelhalter teaches the concept. It can be written without having taken a statistics course, and what you learn writing it will pay off in every departmental course. | Amrhein 2019 · Head 2015 · Spiegelhalter 2019 |
| Misleading with a graph | The technique of producing a false impression with correct numbers, and when that is intent and when carelessness. | You can collect the examples yourself (news sites, institutional reports), so you will have produced primary material. Huff and Tufte give the criteria, Bergstrom-West current examples. | Huff 1954 · Tufte 1983 · Tufte 1990 · Bergstrom 2020 |
| From correlation to causation: how do you get there | What can be said after the sentence 'correlation is not causation'? The focus is not the cliché but the method. | Pearl's book explains causal inference in plain language, Spiegelhalter and Ellenberg give you the habit of statistical reading. A student who writes this topic writes all their other papers better too. | Pearl 2018 · Spiegelhalter 2019 · Ellenberg 2014 |
| Misreading risk | Which risks do people systematically exaggerate, and is that pattern itself predictable? | Kahneman and Ariely give experimental evidence on the same phenomenon; Nudge turns it into policy. The tension between the three (a malfunction or an adaptation) becomes the backbone of the paper. | Kahneman 2011 · Ariely 2008 · Thaler 2008 |
| How good is expert forecasting | Does expertise improve forecast accuracy? Tetlock's answer is uncomfortable. | There is a long term, measured data set, one of the few social science topics where you can say 'we measured it'. Silver does the same job on election and sports forecasts. | Tetlock 2005 · Silver 2012 · Kahneman 2011 |
| Scientific revolution or accumulation | Is science a smoothly accumulating process, or does it advance through breaks? And what does that question say to today's replication argument? | Applying Kuhn's frame to a current phenomenon risks staying abstract, but Ioannidis and Munafò give you concrete material. Combining philosophy with empirical literature is a distinctive move. | Kuhn 1962 · Ioannidis 2005 · Munafò 2017 |
| How climate skepticism was manufactured | With a scientific consensus in place, how does uncertainty get produced in public opinion, and can that be measured? | Oreskes and Conway show the mechanism with archival documents; Wallace-Wells and Kolbert describe the consequence. A good exercise that joins the question 'who paid for it' to source evaluation. | Oreskes 2010 · Wallace-Wells 2019 · Kolbert 2014 |
| Extinction: scale and measurement | Is the 'sixth extinction' a metaphor, or a measured phenomenon? | Kolbert describes the phenomenon, the planetary boundaries studies quantify the biodiversity threshold. Finding the number behind the metaphor can be the paper itself. | Kolbert 2014 · Steffen 2015 · Richardson 2023 |
| Research ethics and consent | Who does a cell line, a data set or a sample 'belong to'? How informative is informed consent in practice? | The case Skloot tells is the classic example of the consent argument; Mukherjee gives the scientific context, Ioannidis the structural failure in the production of evidence. You can use your own department's ethics committee form as primary material. | Skloot 2010 · Mukherjee 2016 · Ioannidis 2005 |
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| Unpublished drug trials | When trials with negative results go unpublished, where does the average of the literature shift? | Easterbrook measures the phenomenon in 1991, Turner shows how much the effect inflates for antidepressants, Anderson measures what changed after registration became compulsory. The counterargument is ready made: did the registration system work? | Easterbrook 1991 · Turner 2008 · Anderson 2015 |
| Inequality in health | Is the life expectancy gap between two neighborhoods in the same city explained by medicine, or by income? | The WHO report gives the data, Marmot the mechanism, Chetty the number for the US. One of the few topics where you will not struggle to find numbers. | Marmot 2005 · WHO 2008 · Chetty 2016 · Wilkinson 2009 |
| Income and life expectancy | Is the relationship between income gaps and life expectancy gaps the same everywhere? What do local differences say? | Chetty's JAMA study measures at both national and local level, which opens up the question 'geography or income' empirically. Wilkinson-Pickett give the comparative claim and have been criticized, so the argument is live. | Chetty 2016 · Wilkinson 2009 · Marmot 2005 |
| How far medicine goes at the end of life | When do 'more treatment' and 'better outcome' come apart? | An emotional topic, which is what makes it hard: write a paper that walks on evidence and you stand out. The risk of sliding into anecdote is high, be aware of it. | Gawande 2014 · Kübler-Ross 1969 · World Health Organization |
| Vaccine hesitancy | A lack of information or a lack of trust? The two point to different policies. | MacDonald gives the definition and the determinants, Horne shows experimentally whether an intervention worked. A thesis showing why the reflex of 'let us give more information' is not enough is both contestable and provable. | MacDonald 2015 · Horne 2015 · World Health Organization |
| How the history of a disease gets written | What kind of knowledge does the 'biography' of a disease produce, medical history or institutional history? | Mukherjee gives the narrative, Ioannidis the failure in the production of evidence, Turner publication bias. If you study molecular biology or psychology you can take the production of knowledge in your own field as the topic. | Mukherjee 2010 · Ioannidis 2005 · Turner 2008 |
| Gene editing: who decides | CRISPR is a technical possibility. Who draws the line, the scientist, the state, or the market? | Doudna, as the inventor of the technology, sets out her own argument about limits, Mukherjee gives historical depth, Skloot the question of consent. If you study molecular biology or genetics it is from your own field, and all three are in the catalog. | Doudna 2017 · Mukherjee 2016 · Skloot 2010 |
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| Climate: the gap between knowing and acting | If the science is clear, why does the decision not change? The focus is not the physics but the decision mechanism. | The IPCC report closes the 'known' part, and the argument lives in what is left. A topic where you will have no problem with source quality. | IPCC 2023 · Klein 2014 · Burke 2015 |
| How the economic cost of climate is calculated | Where does the 'social cost of carbon' come from as a number, and why does it change from country to country? | Burke measures the temperature and output relationship in a non linear form, Ricke brings the same cost down to country level. Showing how a number is manufactured is a very strong paper structure. | Burke 2015 · Ricke 2018 · IPCC 2023 |
| Does carbon pricing work | How much of the emission decline can be attributed to pricing, and why do these estimates diverge? | Andersson measures a single country (Sweden) as a case, Best gives multi country evidence. Arguing about the difference between a single case and cross sectional evidence is the methods lesson itself. | Andersson 2019 · Best 2020 · Ricke 2018 |
| The physical limits of the energy transition | Measuring the distance between the targets and the reality of materials and infrastructure. | Smil proceeds with numbers, the IEA data is current. The less crowded side of the climate topic. Note: the IEA report is annual, so state which edition you used in the citation. | Smil 2022 · Smil 2017 · International Energy Agency |
| Planetary boundaries and growth | Is the concept of a 'boundary' itself measurable, or is it a political choice? | Rockström's 2009 frame gives nine thresholds, and the 2015 and 2023 versions update the same frame, so you can watch how a concept developed over fifteen years. Raworth ties it to economics. | Rockström 2009 · Steffen 2015 · Richardson 2023 · Raworth 2017 |
| The food system and individual choice | Between consumer choice and the supply chain, which one determines the other? | FAO data gives the global picture, Smil the energy and materials side, Meadows the language of systems. The macro to micro move gives the paper a ready structure. | FAO · Smil 2022 · Meadows 2008 |
| Obligations to future generations | Can we owe something to people not yet born, and what does that change in climate policy? | Parfit sets the problem up philosophically (and finds his own paradoxes), Ricke and the IPCC give the numerical side. If you study philosophy it is from your own field; the hardest and most instructive form of tying a value claim to a number. | Parfit 1984 · Ricke 2018 · IPCC 2023 |
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| The place of the car in the city | Does density really reduce car use, and how is that measured? | Ewing and Cervero's meta-analysis gives the relationship between the built environment and travel behavior in numbers, and the 2017 piece is a reply to criticism. So you can see two rounds of the argument together. Jacobs gives the conceptual frame. | Ewing 2010 · Ewing 2017 · Jacobs 1961 |
| Congestion charging | Does traffic pricing work, and by which criterion is 'working' defined? | Leape's JEP piece gives the London scheme with its measured results, open access and short. A clean case showing that the choice of criterion determines the result. | Leape 2006 · Ewing 2017 · Jacobs 1961 |
| The measurement problem in the pedestrianization argument | Is the shopkeepers' objection irrational, or does it come out of the wrong measurement? The question: how does the customer arrive, and who measures that? | The measurement question is clear and your sources argue exactly that: can the relationship between the built environment and behavior be measured, and by which criterion. Exactly the right scale for a first year paper. | Ewing 2010 · Leape 2006 · Jacobs 1961 · Wheelan 2013 |
| Housing, eviction and poverty | Is eviction the consequence of poverty or its cause? A question of direction. | Desmond's field work makes the direction of causality arguable; Harvey gives the theoretical counter reading. | Desmond 2016 · Harvey 2012 · Wilkinson 2009 |
| Short term rentals and the city | Do Airbnb type platforms really affect the housing market, and through which channel? | Zervas measures the effect on the hotel sector, Goldfarb-Tucker give which economic properties of digital platforms are different, Desmond the housing side. If you carry it into the Turkish context you will have to find a data source separately; say so in your thesis as a limit. | Zervas 2017 · Goldfarb 2019 · Desmond 2016 |
| Loneliness and public space | Is the loosening of social ties an individual choice, or the loss of infrastructure? | Putnam sets the problem up, Holt-Lunstad's meta-analysis gives the measured health effect of loneliness, Frumkin the physical environment link. One of the few topics where you can tie a sociological claim to health data. | Putnam 2000 · Holt-Lunstad 2015 · Frumkin 2018 |
| Green space and health | Can the health effect of a park be measured, or are other things already different where the park is? | Frumkin gives the review frame, the WHO report the social determinants. The question itself is a causality question, so it opens straight onto a methods argument. | Frumkin 2018 · WHO 2008 · Jacobs 1961 |
| Air pollution and productivity | The health cost of pollution is known. But how is the productivity cost measured? | Chang and colleagues match the daily output of pear packing workers with pollution data, small scale, transparent in its measurement, a design you could imitate. A good example of method. | Chang 2016 · Burke 2015 · World Health Organization |
| Who dies in a disaster | The same tremor, the same heat wave: why does the death rate change from neighborhood to neighborhood? | Cutter's social vulnerability index gives a measurable frame. If you carry it into the Turkish context, mind the distinction between sources: AFAD's earthquake hazard map gives expected ground motion, not who died, so hazard and harm are two separate things. A paper that sets up that distinction is halfway there. | Cutter 2003 · Ambraseys 2011 · AFAD |
| Earthquake deaths and building control | How much of earthquake deaths is physics and how much is governance? | Ambraseys and Bilham's Nature piece measures the relationship between corruption and earthquake deaths at country level, a peer reviewed and short source that speaks directly to the Turkish context. Cutter gives the vulnerability frame. | Ambraseys 2011 · Cutter 2003 · AFAD |
| Form and meaning in architecture | What measures a building being 'good', formal coherence or use? | Venturi (in the catalog in Turkish translation), Koolhaas and Alexander propose three different criteria and openly contradict each other. If you study architecture or interior architecture it is from your own field, and your counterargument section fills itself. | Venturi 1972 · Koolhaas 1978 · Alexander 1977 |
| How we read the city | How do people map a city in their heads, and is that map data you can use in planning? | Lynch's method (having users draw the city) is a measurable design; Jacobs gives the conceptual counter reading, Alexander translates it into a design language. One of the few topics where you can run a small observation on your own campus. | Lynch 1960 · Jacobs 1961 · Alexander 1977 |
| Experiencing space with the body | Is architecture essentially a visual art? Are the other senses something measurable? | Pallasmaa takes a clear position against ocularcentrism; Berger gives the regime of looking, Lynch the perceptual map. The risk of staying abstract is high, so pick one single space as your case. | Pallasmaa · Berger 1972 · Lynch 1960 |
| Is urban growth good or bad | Does density lower or raise quality of life, and how is 'quality' measured? | Glaeser defends density, Montgomery builds the link between design and wellbeing, Jacobs gives the conceptual ground. All three look at the same phenomenon with a different criterion. | Glaeser 2011 · Montgomery 2013 · Jacobs 1961 |
| Republican architecture and nation building | By which criterion is a building counted as 'modern' or 'national'? | Bozdoğan's two books set this question up in the Turkish context and both are in the catalog; Zürcher gives the historical frame, Anderson the theory of the nation. If you study architecture or history it is directly from your own field, and it is the topic where peer reviewed sources in the Turkish context are easiest to find. | Bozdoğan 2001 · Bozdoğan 1997 · Zürcher · Anderson 1983 |
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| The myth of merit | Does the belief that 'those who deserve it win' reduce inequality, or justify it? | Sandel gives the philosophical claim, Wilkinson-Pickett the comparative data, Arum-Roksa how much actually gets learned at university. One of the few topics where you can tie a value argument to measurement. | Sandel 2020 · Wilkinson 2009 · Arum 2011 |
| What a university is for | Preparation for employment, or the making of citizens? What happens when the two collide? | Taking the institution you are inside as your topic gives you an observation advantage. Arum and Roksa measure the question 'did learning happen', which stops you sliding into personal complaint. The risk: it turning into anecdote, so stay tied to evidence. | Nussbaum 2010 · Deresiewicz 2014 · Arum 2011 · Illich 1971 |
| What you measure goes bad | Once an indicator becomes a target, why does it start measuring the thing it stopped measuring? | Espeland and Sauder track this mechanism empirically in law school rankings, and the concept of 'reactivity' is made for exactly this job. Muller gives the book version (the Turkish edition is in the catalog), O'Neil the algorithmic side. Pick a narrow field and a very good paper comes out. | Espeland 2007 · Muller 2018 · O'Neil 2016 |
| Grade inflation | While grades rise, does learning rise? If the two come apart, what is the grade measuring? | Arum and Roksa measure learning gains directly, Johnson the grade distribution and the effect of grades on course choice (a year long experiment at Duke). Putting the two side by side can be the thesis itself. Your own transcript is not data, so do not use it as data in the paper. I listed Johnson's book and the Academe review written about that book separately: their titles are word for word the same, the review comes up first in a search and its citation gets confused with the book's. The book itself is in the library as an e-book, so go to that. | Johnson 2003 · Perry 2004 · Arum 2011 · Espeland 2007 |
| Lecturing or working them | Has the difference between the classic lecture and active learning been measured, and by how much? | Freeman's meta-analysis combines 225 studies and is open access; Dunlosky collects which study techniques work. The ideal topic against the trap of taking your own classroom experience for evidence, because measured evidence is plentiful. | Freeman 2014 · Dunlosky 2013 · Karpicke 2008 |
| Taking notes: pen or keyboard | Does writing by hand really change learning, and how solid is that finding? | Mueller and Oppenheimer's study is the source of the popular advice; there is also the later replication argument. A topic where you can write both the finding and the fragility of the finding, exactly the right scale for ENG 102. | Mueller 2014 · Wolf 2007 · Dunlosky 2013 |
| How much does 'growth mindset' work | Can a small intervention make a big difference? An argument about effect size. | Yeager's national scale experiment finds that the effect exists but is small and conditional, a study that turns 'does it work' into 'in whom, and how much'. Showing the distance between the popular claim and the measured effect makes a strong thesis. | Yeager 2019 · Dunlosky 2013 · Arum 2011 |
| Choosing a study technique | Why do students trust techniques that do not work, and which technique works as measured? | Dunlosky ranks the techniques by level of evidence, Make It Stick is the book version of the same literature, Dweck and Yeager give the motivation side. A chance to write a paper that changes your own study habits. | Dunlosky 2013 · Brown 2014 · Dweck 2006 · Yeager 2019 |
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| Managing common resources | Under which conditions does Hardin's 'tragedy of the commons' model fail to happen? | Ostrom answers Hardin precisely, and with field examples. Because the two sources talk to each other, the thesis comes out contestable by itself; Meadows gives the language of systems. | Hardin 1968 · Ostrom 1990 · Meadows 2008 |
| Open source and digital commons | How unpaid work sustains itself, a question about institutional structure rather than motivation. | Carrying Ostrom's common resource principles into software joins two fields and is a move your instructor has not seen a hundred times. Lessig gives the legal side. | Benkler 2006 · Lessig 2004 · Ostrom 1990 |
| Refugee policy: camps or labor | On what evidence is the choice between the aid model and the employment model made? | Betts and Collier put forward an explicit policy proposal, so there is a target you can argue against. UNHCR data gives the magnitudes, Standing the labor side. | Betts 2017 · UNHCR · Standing 2011 |
| Nudging and autonomy | Where is 'pushing a person for their own good' help, and where is it manipulation? | Thaler and Sunstein defend it, Misbehaving tells the field's own history, Haidt how moral intuitions work. Demanding because you are forced to tie an ethical claim to empirical evidence, and instructive for the same reason. | Thaler 2008 · Thaler 2015 · Haidt 2012 |
| Whose fault is a design error | When the user uses it 'wrong', who is at fault? The focus is not user training but design responsibility. | Norman gives the conceptual frame, Vaughan the organizational anatomy of a disaster, Catalano responsibility in the language of engineering ethics. If you study engineering or design it is directly from your own field. | Norman · Perrow · Catalano 2006 · Vaughan 1997 |
| Do disasters come out of organizational culture | Where do technical failure and organizational failure come apart? Is 'normalized deviance' something measurable? | Vaughan's Challenger study is the source of that concept and the field material is plentiful. Meadows gives the language of systems, Catalano the ethical frame. | Perrow · Meadows 2008 · Catalano 2006 · Vaughan 1997 |
| How the state sees | Do the failures of central planning come from a lack of information, or from the kind of information? | Scott's frame (the distinction between local knowledge and administrative knowledge) fits both historical and current examples; Why Nations Fail gives the institutional account, Polanyi the market and state relationship. | Scott 1998 · Acemoglu 2012 · Polanyi |
| How far the market goes | Which things lose their quality once they are bought and sold, and is that claim measurable? | Polanyi gives the theoretical ground, Sandel the current moral argument. Hard because you have to tie a value argument to institutional history, but distinctive. | Polanyi · Sandel 2020 · Acemoglu 2012 |
| What punishment achieves | Does imprisonment reduce crime, or is it managing something else? | Alexander gives the structural claim, Stevenson the field narrative, Rawls the normative criterion. If you study law it is from your own field; if not, it is the most demanding form of the 'tie a value claim to evidence' exercise. | Alexander 2010 · Stevenson 2014 · Rawls 1971 |
| How a criterion of justice gets built | By which criterion can we say a distribution is 'just'? | Rawls gives the frame, Sandel the objection, Wilkinson-Pickett the empirical picture. Hard because you are forced to bridge philosophy and data, but it stands out if written well. | Rawls 1971 · Sandel 2020 · Wilkinson 2009 |
| Information empires: the cycle that opens and closes | Why do communication technologies start open and then become monopolies? Does Wu's cycle claim hold today? | Wu proposes an explicit historical pattern, so it is falsifiable; Lessig gives that architecture sets the rules, Vaidhyanathan the current case. All three are in the catalog. | Wu 2010 · Lessig 2006 · Vaidhyanathan 2018 |
| Entry | The angle you take | Why it is arguable | Sources |
|---|---|---|---|
| Looking at pain: photography and witnessing | Does an image of suffering move people to act, or desensitize them, and can that be measured? | Sontag gives a different answer to the same question in 1977 and in 2003; the second book openly takes back some claims of the first. Tracking a writer who changed her own mind is a strong paper structure, but you have to open both books. | Sontag 1977 · Sontag 2003 · Berger 1972 |
| Machine translation and translation quality | How is translation quality measured, and has machine translation changed those criteria? | House gives the theoretical frame of quality assessment, Li the practice of post-editing. If you study translation and interpreting it is from your own field; if not, it is a clean example of 'how a criterion gets built'. | House · Li 2023 · Mishra 2018 |
| Language loss and bilingualism | The loss of a language is the loss of what? Turning the answer 'cultural richness' into a measurable claim. | Crystal gives the frame (the book is in the catalog, and also as an e-book on Cambridge Core); Mishra the cognitive side of bilingualism. Careful: UNESCO's interactive atlas of endangered languages was shut down and the atlas that replaced it is offline too, so if you are after a current number do not rely on a single source, and state its date. | Crystal 2000 · Mishra 2018 · House |
| Museums and the restitution of cultural property | Who does an artifact 'belong to'? The focus is not emotional ownership but the legal and institutional criterion. | Cuno and Merryman set up two opposing positions clearly (cultural nationalism and cultural internationalism), so your counterargument section arrives ready made. If you study archaeology or art history it is from your own field; if not, Berger gives the regime of looking side. | Merryman 2005 · Renfrew · Berger 1972 · Cuno 2010 |
| The economics of mega events | Does the promised economic return of hosting an Olympics or a World Cup actually materialize? | The economists' answer is largely clear but public belief runs the other way, and that gap is itself a researchable question. Baade and Matheson are open access and short. | Zimbalist 2015 · Baade 2016 · Zimbalist 2002 |
| Refereeing decisions and social pressure | Are referees' decisions affected by crowd pressure, and how do you measure that cleanly? | Garicano and colleagues measure it through added time; a textbook example of natural experiment design. The way to turn the argument on the five a side pitch into an academic question. | Garicano 2005 · Zimbalist 2002 · Kahneman 2011 |
| How visual narrative works | A comic, an infographic or a data chart: how does a visual narrative build meaning? | McCloud gives the grammar, Tufte the criterion, Berger the regime of looking. If you study graphic design, communication or fine arts it is from your own field, and all the sources are in the catalog. | McCloud 1993 · Tufte 1990 · Berger 1972 |
| What a photograph does | Is a photograph evidence, or interpretation? Barthes and Sontag answer the same question differently. | Barthes argues about photography's own kind of time, Sontag about its moral effect. Both are in the catalog (Camera Lucida in Turkish translation). If you work on visual communication or photography it is directly from your own field. | Barthes 1980 · Sontag 1977 · Berger 1972 |
| Do everyday objects produce myth | Does the reading Barthes applied to advertising in 1957 fit today's brand communication? | Applying an old method to new material is both original and controlled. You have to show which part of the method carries over, and that sharpens the thesis. | Barthes 1957 · Berger 1972 · McCloud 1993 |
| Whose narrative is the narrative of art history | What does an 'art history' book get built by leaving out? | Gombrich is the canon itself, Said the theory of the mechanism of leaving out, Berger the regime of looking. Because the three sources are in tension, the thesis comes out contestable by itself. | Gombrich · Said 1978 · Berger 1972 |
| What music does in the brain | Is the cognitive effect of music something measurable, or is it storytelling? | Sacks is case based, Levitin collects the experimental literature, and the difference in method between them (case study versus group data) can sit at the center of the paper. If you study music or psychology it is from your own field. | Sacks 2007 · Levitin 2006 · Kahneman 2011 |
| Deciding by numbers in sport | How far can a data driven decision beat intuition, and does that turn into winning? | Moneyball gives the narrative, Zimbalist the theory of competitive balance, Garicano a measured bias. A good footing for learning to build a numerical argument. | Lewis 2003 · Zimbalist 2002 · Garicano 2005 |
| How a nation is imagined | Is nationalism an old feeling or a modern technology? Anderson's answer points at print. | Anderson's claim proposes an openly falsifiable mechanism (print publication + language standardization); Said gives the side of leaving out, Herman-Chomsky the media model. If you study international relations or political science it is from your own field. | Anderson 1983 · Said 1978 · Herman 1988 |
| Does language shape thought | Which part of the Sapir-Whorf claim survived empirically? | Deutscher defends the weak version of the claim, Pinker argues against the strong one, Mishra brings the bilingualism data. Tracking how a thesis got narrowed over time is a strong paper structure. | Deutscher 2010 · Pinker 1994 · Mishra 2018 |
| How meaning gets built in film | Is editing a technique, or a way of building an argument? | Bordwell gives the language of analysis, Murch editing's own logic, Berger the regime of looking. If you study communication or design it is from your own field; picking a single scene as your case keeps the scope under control. | Bordwell · Murch · Berger 1972 |
| What a game measures | Does gamification raise motivation, or does it just make measuring easier? | Huizinga (in Turkish translation in the catalog) gives the cultural function of play, Juul the rules and fiction distinction, McGonigal and Bogost the design claims. If there is a gamified application in your own department, primary observation material is at hand. | Huizinga 1938 · Juul 2005 · McGonigal 2011 · Bogost 2007 |
| Typography and readability | Is the readability of a text something measurable, or a matter of stylistic preference? | Lupton gives the design criteria, Tufte the criterion of information density, McCloud the visual grammar. If you study graphic design it is from your own field, and all three are in the catalog. | Lupton · Tufte 1990 · McCloud 1993 |
| Method and interpretation in archaeology | How much can be said about a 'society' from a find? The distance between evidence and interpretation. | Renfrew and Bahn are the standard reference on method; Cuno and Merryman give the argument about who cultural property belongs to. If you study archaeology it is from your own field. | Renfrew · Merryman 2005 · Berger 1972 |
| Why twentieth century music got difficult | Is a piece of music being counted 'unintelligible' in the listener, in the work, or in the institution? | Ross gives the historical narrative, Sacks and Levitin the perception side. If you study music it is from your own field; if not, it is a good example of the question 'can taste be measured'. | Ross 2007 · Sacks 2007 · Levitin 2006 |