Skip to main content

Bias in Social Sciences Research

By January 15, 2025Commentary

Illegal immigration is a serious problem in this country, leading to crime, important of deadly drugs, potential terrorist infiltration and high costs of supporting those here illegally.  Legal immigration has been extremely beneficial to our country, bringing in skilled workers.  But even legal immigrants can generate high costs initially.  There has been a large body of research on the relative economic costs and benefits of immigration.  Some of it is done by people with a blatant point of view; some purports to be neutral and objective.

There is a serious problem in the “social” “sciences” about the credibility of research.  This is generally softer research not dealing with physical problems.  It is harder to replicate a particular study and its outcomes, and replication rarely occurs.  So the results are easy to manipulate or even make up and that happens frequently.  Even in supposed “harder” sciences, like climatology, there is clear evidence of data manipulation, poor methodologies, and biased results.  Science has a crisis of credibility because scientists far too often stray into advocacy and politics.  And then they wonder why trust in science is at all-time lows.

This paper looks at ideological bias in research on the question of the costs and benefits of immigration.  158 researchers on 73 teams were given the same data on whether it appeared that immigration affected public support for social welfare programs.  These researchers had all been surveyed on their attitudes toward immigration before doing the research.  Gee, guess what, the researchers who had pro-immigration views were much more likely to find that immigration increased support for social programs and those who less strongly favored immigration were more likely to find a negative effect.  The differing results were reached by making different design and methodology choices and those researchers who had findings futher from the average result were more likely to use questionable designs and methodology.   (NBER Paper)

The lesson should be that researchers are required to fully disclose all political affiliations and views, must show what the results of their research would be across a range of design and methodology choices and associate themselves with researchers who have opposing views.  It is critically important that the public trust research and data and analysis, and that won’t happen if scientists and researchers don’t start being more transparent and frankly, question themselves about their biases and the impact of those biases on the work they produce.

Join the discussion One Comment

  • Joe K says:

    I see tremendous amount of bias in much of politically charged “medical Science”

    Lots of junk science / borderline academic fraud level studies that are highly touted, yet close examination would reveal the substandard nature of the work

    To name a few
    The gas stove causing 12% of asthma cases study using the inappropriate Population attribution fraction methodology,
    The study showing higher excess death among republicans than democrats in Ohio and Florida = especially absurd when data of deaths by party affiliation is completely unreliable.
    The Kansas mask mandate vs no mask mandate by county study in which the gap completely closed when the study period was cut short.

    One of the best examples is a 2000 study of increased premature mortality in 96 US cities when there is an 10ppm increase in ground level ozone (Bell McDerment). Lots of biases in the data and in the conclusion, missing the higher factor for the cause of increased death rates.

    In the ozone study, it doesnt appear the study authors were trying to commit academic fraud, though they did let biases affect their conclusions.

    With the first three studies, it definitely appears to be outright academic fraud, especially the asthma study and the excess death study.

Leave a comment