Experimental Methods Flashcards Preview

MCAT Sociology > Experimental Methods > Flashcards

Flashcards in Experimental Methods Deck (51)
Loading flashcards...
1
Q

Studies can fall into two categories based on the role of the researcher/experimenter. Name these two categories.

A

Experimental and observational studies

As their names indicate, experimental studies involve the researcher manipulating some condition(s) experienced by the groups, while in observational studies, the researcher simply observes and/or measures the sample without interfering.

2
Q

Do randomized controlled trials (RCTs) exemplify observational or experimental studies?

A

Randomized controlled trials (RCTs) exemplify experimental studies.

Both the word “randomized” and the word “trial” are hints to the nature of an RCT. “Randomized” indicates that subjects are randomly assigned to different groups, which occurs only in experimental studies. “Trial,” too, has a connotation of experimenting rather than observing.

3
Q

How large, at minimum, must a sample size be for a study to be valid?

A

There is no minimum sample size, although the larger the sample, the better.

This is important! In studies of extremely rare conditions, for example, it’s not possible to utilize a very large sample, and that’s okay. Be wary of MCAT answer choices that indicate that a study is flawed simply because it has a small sample size, unless the passage/question gives additional information to support this.

4
Q

Observational studies differ from experimental studies in that observational studies cannot prove:

A

causation.

Since observational studies involve only observation, these studies are highly subject to confounding variables. Additionally, even the direction of causation is often unclear in such a study. For these reasons, only experimental studies can determine causation.

5
Q

Name the type of observational study in which a sample is followed and measured over time.

A

A longitudinal study

Longitudinal studies are advantageous in that they allow researchers to determine the sequence of events that subjects experience. Subjects are often followed for months or years.

6
Q

Name the type of observational study in which samples are evaluated once, at a single point in time, and compared.

A

A cross-sectional study

These studies are highly useful in comparing variables (often, many variables) at a time, but they cannot be easily used to determine sequences of events over time.

7
Q

A classic distinction in research is between ________ and ________.

This distinction refers to whether a change in one variable directly leads to a change in another, or when both changes simply happen to be more likely to occur together.

A

A classic distinction in research is between correlation and causation.

Causation, the stronger of the two relationships, occurs when a change in one variable causes a change in another variable.

Correlation occurs when a change in one variable (say, increased blood glucose) simply happens to occur with, but is not necessarily caused by, a change in another variable.

8
Q

Name the value that serves as a statistical measure of the strength of the correlation between two variables.

A

The correlation coefficient

This value ranges from −1.0 to 1.0.

9
Q

True or false:

A negative correlation means that two variables are not associated with each other at all.

A

False

This is incorrect! A negative correlation means that two variables are associated with each other; specifically, as one variable increases, the other decreases.

In contrast, a positive correlation means that as one variable increases, the other also increases.

10
Q

A complete lack of correlation between two variables is indicated by a correlation coefficient that is close to what value?

A

0

A correlation coefficient of 0 indicates a total lack of correlation; in other words, the two variables have no relationship with each other at all.

11
Q

As GPA increases, MCAT score also tends to increase. If a large increase in GPA only produces a relatively small increase in MCAT score, the correlation coefficient of this relationship would likely be between what two values?

A

Between 0 and 0.5

A correlation coefficient in this range indicates the presence of a weak positive correlation (in other words, as one variable increases by a large amount, the other variable increases by a small amount).

12
Q

A strong negative correlation is indicated by a correlation coefficient near what value?

A

−1

The strongest possible negative correlation is indicated by a correlation coefficient of −1.

13
Q

A weak negative correlation is indicated by a correlation coefficient between ________ and ________.

A

A weak negative correlation is indicated by a correlation coefficient between 0 and −​0.5.

A weak negative correlation means that as one variable increases by a large amount, the other variable decreases by a small amount.

14
Q

When two variables are plotted graphically, the resulting best-fit line is linear and steeply sloping upward. This indicates a correlation coefficient close to:

A

1.

This description corresponds to a strong positive correlation (as one variable increases, the other similarly increases). The strongest possible positive correlation is denoted by a correlation coefficient of 1.

15
Q

A correlation matrix is a table in which multiple variables are plotted along both the columns and rows of the table. Each cell then shows the correlation coefficient between the variable listed for that column and that listed for that row. Correlation matrices are typically marked by a line of correlations of 1 running diagonally through them. Why?

A

This diagonal line exists because the correlation between a variable and itself is always 1.

Check this out yourself next time you see a correlation matrix!

16
Q

A longitudinal study in which a group of individuals who share some characteristic are followed over time is termed a:

A

cohort study.

Here, the group of individuals is termed a cohort. Cohorts may be similar in their birth year (a “birth cohort”), in whether they have some disease, in their early exposure to a toxin, or in many other ways.

17
Q

Unlike a cohort study, a ________ study involves close, detailed analysis of an individual, small group, or situation.

A

case study

This differs from a cohort study, which involves analysis of a larger group. Case studies are extensively used in medicine, especially for rare or previously unknown conditions (such as a deep dive into the case of a woman with a previously undiscovered type of tumor).

18
Q

What type of study is conducted at the population or community level and typically focuses on the risk factors for a disease or condition?

A

An ecological study

Often, the population assessed in an ecological study is determined geographically. For instance, an epidemiological study of risk factors for infection by a new virus in South Korea would be an ecological study.

19
Q

Research methods that involve the observation of people or groups in their natural settings are termed:

A

ethnographic methods.

These methods often are used to gather qualitative data about societies, although they can also be used to obtain quantitative data. Ethnographic methods are extensively used by anthropologists and some sociologists.

20
Q

________ is a research method that involves analyzing the words and (sometimes) images from websites, documents, works of art, or other bodies of language.

A

Content analysis

This method has been used increasingly often since the advent of computing tools that allow for easy analysis of text. (If you’ve ever seen a word cloud, that is an example of content analysis.)

21
Q

True or false:

Experimental studies are often a good way to study ecological factors at the neighborhood or community level, such as residential segregation.

A

False

Experimental studies at the neighborhood or community level are often unethical and typically impossible. (For instance, an experimental study of residential segregation would need to place some people in segregated areas and other people in areas that are not segregated, which is both not feasible and unethical.)

22
Q

Define:

survey methods

A

Survey methods are research methods intended to gather large amounts of data from subjects, often via tools such as questionnaires.

23
Q

Name three advantages of survey methods.

A

Survey methods are:

  • relatively cheap to use
  • relatively easy to implement (imagine how much simpler it is to write up a short survey than to design an experimental study!)
  • and can reach a much larger sample size than a typical experimental trial.
24
Q

All but the very largest surveys typically are administered to a smaller ________ with the goal of reflecting the entire ________.

A

All but the very largest surveys typically are administered to a smaller sample with the goal of reflecting the entire population.

The exceptions to this rule are large surveys like the U.S. Census, which aim to evaluate the entire population. (In those cases, then, the sample is the population.)

25
Q

For the results of a survey to properly reflect a population, it is critical that the surveyed sample be:

A

representative.

For instance, if the population we are trying to understand contains both men and women, it would be unrepresentative (and methodologically flawed) to survey only men.

26
Q

True or false:

All survey samples must include individuals from all wealth quintiles (from the wealthiest to the most impoverished).

A

False

This depends entirely on what the survey is aiming to study! For example, a survey focused solely on those below the poverty line would not need to include wealthy individuals to be representative.

27
Q

In a given population, 68% of the heads of household are renters, while 32% are homeowners. In a survey of heads of household with a sample size of 5,000 people, how many should be homeowners?

A

1,600

A perfectly representative sample should mirror the demographics of the population being evaluated. An easy way to obtain this answer is to realize that 32% of 10,000 would be 3,200, so 32% of 5,000 must be half of that: 1,600.

28
Q

The concept in research that posits that understanding causal relationships requires a sense of which events happen before others is termed:

A

temporality.

This concept follows from the logical idea that a cause must always precede its effect.

29
Q

Define:

a questionnaire

A

A questionnaire is a list of questions devised by researchers and administered to or sent to a sample of people.

30
Q

What is self-reporting bias?

A

Self-reporting bias, or response bias, is a general term for a set of biases that cause survey subjects to respond inaccurately to survey questions.

Examples of self-reporting bias include social desirability bias and acquiescence bias.

31
Q

The tendency to agree with a statement when one is unsure is termed:

A

acquiescence bias.

You’ve probably experienced this before: you’re asked a question that you aren’t fully paying attention to or don’t really care about, and you reflexively say “yeah, sure.” This is a major problem that must be accounted for in surveys.

32
Q

How can researchers mitigate the effect of acquiescence bias in a questionnaire?

A

Include questions that have opposite meanings

Acquiescence bias is the tendency to answer “yes” when uncertain or uninvested in a question. If questions with opposite meanings are asked, participants who answer “yes” to both can be identified, and those results can be thrown out or otherwise accounted for.

33
Q

A type of self-reporting bias that is particularly frequent when individuals are asked about their salaries is:

A

social desirability bias.

This is the tendency to answer questions in a way that make us look or feel better about ourselves. While salary is a classic example, this bias appears in questions about everything from one’s height to social and political issues.

34
Q

A person is answering a survey about a political candidate. The only thing the person knows about the candidate is that he is well-dressed and speaks coherently. The survey respondent therefore assumes that the candidate is a good person. This exemplifies what bias?

A

The halo effect

The halo effect refers to a cognitive bias in which a person extrapolates from positive impressions about one aspect of a person or situation, usually resulting in an assumption that all facets of that person or situation are good/positive.

35
Q

What is the difference between leniency bias and harshness bias?

A

Leniency bias refers to the tendency to answer every question leniently (that is, to be “too nice” in one’s answers).

Harshness bias refers to the opposite: the tendency to answer every question harshly, or be “too mean” in one’s answers.

36
Q

This type of scale, used extensively in surveys, asks respondents to rate their level of agreement or disagreement with an item.

A

A Likert scale

For example, a 5-item Likert scale may consist of the responses “strongly disagree,” “disagree,” “neither agree nor disagree,” “agree,” and “strongly agree.”

37
Q

What term in epidemiology refers to the number of new cases of a disease per unit of population over time?

A

Incidence

For instance, the incidence of a particular cancer may be an average of 3 new cases per 10,000 people annually.

Note that epidemiology refers to the study of disease and its spread.

38
Q

What term in epidemiology refers to the proportion of people who have a given condition or disease at a given time?

A

Prevalence

For instance, the current prevalence of a particular cancer may be 23 people who have the cancer for every 10,000 people.

39
Q

Which value is likely to be higher: the annual incidence rate of human immunodeficiency virus (HIV) or the prevalence rate of HIV?

A

The prevalence rate of HIV

Since the incidence rate refers only to new cases of HIV developed over the previous year, while the prevalence rate includes everyone with HIV (even those who acquired the virus prior to the previous year), the prevalence rate should be higher.

40
Q

Studies of what population have been integral to our understanding of the classic “nature vs. nurture” debate?

A

Twins

Twin studies have been absolutely crucial to our understanding of the roles played by nature (genetics) versus nurture (one’s environment). Since identical twins share the same genetics, twins raised in different environments allow us to isolate the effects of the “nurture” side.

41
Q

Monozygotic twins share essentially ____% of their genes.

A

100%

Monozygotic, or single-zygote, twins are also known as identical twins. Apart from the influence of rare phenomena such as mutations during development, these twins share an identical genetic code.

42
Q

Dizygotic twins share ____% of their genes.

A

50%

Dizygotic, or double-zygote, twins are also known as fraternal twins. Genetically, these twins are no different from any non-twin siblings: they share approximately 50% of the same genes.

43
Q

True or false:

If twins who were adopted into different families exhibit dramatically different interests, this supports the “nurture” side of the nature vs. nurture debate.

A

True

In adoption studies such as the one mentioned here, differences between the twins constitute evidence of “nurture” (the impact of environment rather than genetics).

44
Q

Name the three fundamental principles of research ethics.

A
  • Respect for persons
  • Beneficence
  • Justice

The rules of research ethics (such as the need to obtain informed consent) are all intended to uphold these three principles.

45
Q

Describe the ethical principle of beneficence in research.

A

Beneficence is the principle of maximizing the good that is done and minimizing the harm to participants.

This does not mean that participants cannot be harmed at all, but rather that any harm done must be as minimal as possible to achieve the planned benefits of the study.

46
Q

Describe the ethical principle of justice in research.

A

Justice refers to the obligation to distribute the burdens and benefits of research fairly.

Regarding burdens, this means that it would be unethical to conduct a study of a new medication on, say, orphans just because orphans are easier to exploit.

Regarding benefits, it means that the information learned from research should not be hoarded by any group (for instance, the rich).

47
Q

In the infamous Tuskegee Syphilis Study, 600 black men (many impoverished and over half with untreated syphilis) were monitored for over 40 years while being misled about the actual purpose of the study. Even when penicillin was found to effectively treat syphilis, these men were still not given penicillin. Which principle(s) of research ethics did this violate?

A

All three principles

The Tuskegee study is one of the most egregious examples of a lack of research ethics in modern American history. It violated respect for persons (by not obtaining informed consent), beneficence (by not minimizing harm to participants), and justice (by focusing on a vulnerable population).

48
Q

True or false:

Ethical approval committees never approve proposed research studies that involve inflicting pain on subjects.

A

False

Studies are absolutely allowed to inflict pain as long as 1) it is the minimum amount of pain necessary to achieve the study’s goals and 2) the subjects provide informed consent.

(In fact, Clara, the author of this card, once participated in a study that involved shocking participants inside an fMRI machine. That was a terrible time, but $20 is $20, I guess.)

49
Q

True or false:

It is unethical for researchers to include subjects with psychological disorders in their study samples.

A

False

Research studies include subjects with psychological disorders all the time! Again, it is simply extremely important that these subjects be able to provide informed consent.

50
Q

Research involving human subjects can be made more difficult by the ________, which is the principle that people often change their behavior when they know they are being observed.

A

Hawthorne effect

This effect was first noticed when observing workers in a factory for research involving productivity. Predictably, these workers worked differently (harder) because they knew they were being watched.

51
Q

Would you like to give us feedback on this class?

There might be a gift involved.

A

Fill out this survey for a chance at winning a $50 Amazon gift card.

Thank you! We appreciate you helping us create even better flashcards.