This article summarises a research paper by Xiangdong (Tony) Wang of Weston High School in Weston, Massachusetts. It reports the paper''s methods and findings; the full manuscript, tables and code supplement are the author''s own.

A stock can be a serious investment and an internet trend at the same time. In the 2024 FINRA Investor Survey, 29 percent of investors aged 18 to 34 reported buying a meme stock or another investment trending on social media. In the same age group, 50 percent said that entertainment, excitement, fun or playing a game described why they invest, yet 95 percent also said they invest to make money over the long term. Enjoyment and long-term goals clearly coexist.

The paper asks a sharper question. Are young investors who see investing as entertainment more likely to have bought a meme or trending investment, even after accounting for other things that differ between them: how much they know, how much risk they will accept, how long they have been investing and how large their portfolio is?

The data and the design

The study uses the 2024 FINRA National Financial Capability Study Investor Survey, collected online from July to December 2024, and focuses on the 213 respondents aged 18 to 34. After excluding incomplete answers, the main model covers 183 people, 58 of whom reported a meme or trending purchase. The author uses weighted logistic regression and treats the entertainment motive as three categories (not at all, somewhat, very well). Objective investment knowledge is measured by a ten-question quiz, not by how knowledgeable respondents say they are, because FINRA''s own research found that social-media-reliant investors score lower on the quiz while rating themselves more highly.

The paper tests three hypotheses: that entertainment-motivated investors have higher odds of a meme or trending purchase (H1), that higher objective knowledge weakens that link (H2), and that entertainment is linked to frequent trading (H3).

What the data show

The raw pattern is clear. The weighted purchase rate rises from 21.9 percent among those who say entertainment does not describe their motive at all, to 37.5 percent for "somewhat" and 45.1 percent for "very well." Those who see investing as entertainment also report greater risk willingness, less investing experience and heavier reliance on social-media groups and popular investments shown in mobile apps.

Once those traits are controlled for, the picture changes. The odds ratios fall to 1.50 (somewhat) and 1.84 (very well), and they are no longer statistically significant (joint p = .386). The confidence interval for the "very well" group runs from 0.76 to 4.46, which is compatible with anything from somewhat lower odds to a large positive association. The author therefore calls H1 inconclusive, and is explicit that a non-significant result is not evidence that entertainment does not matter.

H2 is not supported: knowledge does not clearly weaken the entertainment link (joint p = .093), and the pattern does not move in the predicted direction. H3 is also not supported (p = .542), though only 27 respondents were frequent traders, so this result is exploratory.

Knowledge: higher, but still low

Purchasers scored higher on the quiz than non-purchasers, averaging 4.31 correct answers out of ten against 3.68. The paper cautions against reading too much into this. Both groups got fewer than half the questions right, so the figures do not show that purchasers are well informed. The positive knowledge coefficient in the main model is not consistently significant across the robustness checks.

What the study cannot tell us

The author is careful about the limits of the evidence:

  • Cross-sectional data. The survey cannot show whether entertainment, risk appetite, experience or the purchase came first. If enjoying investing raises risk willingness, controlling for risk willingness removes part of the very effect being studied.
  • A broad outcome. The survey asks whether a respondent has ever bought a meme stock or other trending investment. It records no asset, date, position size or return, so the findings say nothing about gambling-like behaviour or investment performance.
  • A small subsample. With 58 purchasers in the main model, the confidence intervals are wide, and the results are approximations, not precise estimates.
  • Self-reports. Answers were not checked against brokerage records.

The paper concludes that the adjusted results describe conditional associations, not a total effect of an entertainment motive, and that more definitive answers would need larger samples with motives measured before purchases, along with brokerage records and purchase dates.

The author thanks Bin Shi, Lecturer in Finance at Boston University''s Questrom School of Business, for reviewing the research proposal and manuscript.

Sources

  • Ai, C. and Norton, E. C. (2003), "Interaction terms in logit and probit models," Economics Letters, 80(1), 123–129 — doi.org
  • Chapkovski, P., Khapko, M. and Zoican, M. (2026), "Trading gamification and investor behavior," Management Science, 72(1), 32–56 — doi.org
  • Dorn, D. and Sengmueller, P. (2009), "Trading as entertainment?" Management Science, 55(4), 591–603 — doi.org
  • FINRA Investor Education Foundation (2025), 2024 NFCS Investor Survey instrument — finrafoundation.org
  • FINRA Investor Education Foundation (2025), 2024 National Financial Capability Study: Investor Survey methodology — finrafoundation.org
  • FINRA Investor Education Foundation (2025), National Financial Capability Study terms of use — finrafoundation.org
  • FINRA Investor Education Foundation, NFCS data and downloads — finrafoundation.org
  • Ko, K. J., Green, S., Mottola, G. R. and Valdes, O. (2026), Finfluencer followers and social media scrollers, FINRA Investor Education Foundation — finrafoundation.org
  • Lin, J. T. et al. (2025), Investors in the United States: A report of the National Financial Capability Study, FINRA Investor Education Foundation — finrafoundation.org
  • Pandey, I. and Guillemette, M. (2025), "Social media, investment knowledge, and meme stock trading," Journal of Behavioral Finance, 26(4), 520–536 — doi.org
  • Philander, K. S. (2023), "Meme asset wagering: Perceptions of risk, overconfidence, and gambling problems," Addictive Behaviors, 137, 107532 — doi.org
  • Reiter, M., Qing, D. and Nations, M. (2023), "Who uses social media for investment advice?" Journal of Financial Planning, 36(9), 78–99 — financialplanningassociation.org
  • Vittinghoff, E. and McCulloch, C. E. (2007), "Relaxing the rule of ten events per variable in logistic and Cox regression," American Journal of Epidemiology, 165(6), 710–718 — doi.org