Mid term 2 reading

What is measurement in psychology?

Measurement in psychology involves assigning scores to individuals in a way that is meant to accurately represent a psychological characteristic, such as intelligence, depression, or self-esteem. Researchers must test whether those scores actually reflect the intended construct.

How is test-retest reliability typically assessed?

It is assessed using a correlation, usually Pearson’s r, between scores from time 1 and time 2. A correlation of about +.80 or higher is typically considered good.

What is Cronbach’s alpha?

Cronbach’s alpha is a statistic used to estimate internal consistency. It reflects how closely related the items in a scale are and is often interpreted as the proportion of variance attributable to the true score.

What value of Cronbach’s alpha is usually considered adequate?

A Cronbach’s alpha of about .80 or greater is often considered good, although some sources suggest .70 as a minimum acceptable threshold depending on the context.

What does a negative Cronbach’s alpha suggest?

A negative alpha usually indicates a problem, such as negatively correlated items, poorly coded reverse items, or items that do not belong together conceptually.

What is interrater reliability?

Interrater reliability is the consistency of judgments made by different observers or raters assessing the same behaviour or characteristic.

How is interrater reliability assessed?

It can be assessed using Cronbach’s alpha when ratings are quantitative or Cohen’s kappa when ratings are categorical.

What does perfect reliability mean?

Perfect reliability means that all observed score variance is true score variance and none is due to error. In practice, this is almost never achieved.

How can reliability be defined in terms of variance?

Reliability can be defined as the proportion of total observed variance that is due to true score variance rather than error variance.

What is the relationship between scale length and reliability?

Longer scales tend to be more reliable because having more relevant items generally increases internal consistency. However, longer scales also increase participant burden.

Why can short scales be problematic?

Short scales are efficient, but they may have lower reliability because fewer items provide less information and may not capture the construct as fully.

How can researchers improve reliability?

Researchers can improve reliability by increasing the number of good items, improving item quality, ensuring items assess the same construct, and reducing sources of measurement error.

What is discriminant validity?

Discriminant validity is shown when a measure does not correlate too strongly with measures of conceptually different constructs.

Why is discriminant validity important?

It shows that a measure is specific and not just reflecting general distress, mood, or some other unrelated factor.

What kinds of evidence are relevant when evaluating a measure?

Relevant evidence includes test-retest reliability, internal consistency, interrater reliability, face validity, content validity, criterion validity, convergent validity, and discriminant validity.

Why are reliability and validity ongoing processes?

Neither reliability nor validity is established once and for all in a single study. Researchers continue gathering evidence across samples, settings, and uses to determine whether a measure remains appropriate.

What is statistical power?

Statistical power is the likelihood of detecting a real effect or difference when one truly exists.

How does reliability affect statistical power?

Higher reliability increases statistical power because it reduces measurement error. This makes true relationships easier to detect.

Why can improving reliability be similar to increasing sample size?

Both improving reliability and increasing sample size reduce the impact of error and make it easier to detect real effects. A more reliable scale can sometimes achieve the same power as a larger sample.

How can low reliability weaken a study?

Low reliability increases error variance, reduces power, weakens correlations, and makes it harder to detect real differences between groups or relationships between variables.

What sample size did Nunnally suggest for scale development?

Nunnally suggested around 300 subjects as a useful benchmark, although acceptable sample size depends on the number of items and the complexity of the scale.

What is quantitative nonrepresentativeness?

Quantitative nonrepresentativeness occurs when the sample has a narrower or different range of the construct than the target population, making estimates less precise.

What is qualitative nonrepresentativeness?

Qualitative nonrepresentativeness occurs when participants interpret items differently from the intended population, often due to culture, language, or lived experience.

Why is qualitative nonrepresentativeness especially problematic?

Because it can change the meaning of items and distort the structure of the measure, making the scale less valid for the population it is supposed to assess.

Why are Internet surveys increasingly popular?

Internet surveys are inexpensive, efficient, and can reach many people quickly. Research also suggests that their findings often align with those from more traditional methods.

What resources can help researchers find existing measures?

Mental Measurements Yearbook, Tests in Print, and online initiatives like PROMIS.

What is research ethics?

Research ethics refers to the moral principles that guide how research is designed, conducted, and reported, especially when human participants are involved.

Why is research ethics so important in psychology?

Because psychological research can affect participants’ well-being, privacy, autonomy, and trust. Ethical standards protect participants, the scientific community, and society.

What was problematic about the 1998 Lancet article linking the MMR vaccine to autism?

The study had serious flaws, including conflict of interest, biased participant selection, unapproved and unnecessary procedures, and conclusions that could not be replicated. It was later retracted, and the lead researcher lost his medical licence.

Why is the MMR-autism case important in research ethics?

It shows that unethical or flawed research can create widespread public harm, mislead society, and undermine trust in science.

What did Diederik Stapel do?

Stapel fabricated data in dozens of studies, damaging the credibility of research, harming collaborators and students, and violating the trust that science depends on.

What did Michael LaCour do?

Michael LaCour fabricated research findings in political psychology, which led to a major loss of credibility and professional consequences.

What is respect for persons?

Respect for persons means recognizing the autonomy of participants, ensuring informed consent, and providing extra protection for vulnerable people who may not be able to fully protect their own interests.

Why is justice especially important in research?

Because historically marginalized or vulnerable groups have often been exploited in research, so ethical research must ensure fairness and equitable treatment.

Why is Milgram’s obedience study ethically controversial?

Although it produced important findings, participants experienced significant distress, believed they were harming others, and often required debriefing afterward.

What is the Nuremberg Code?

The Nuremberg Code was created after World War II and emphasized voluntary informed consent and careful risk-benefit evaluation in human research.

What is the Declaration of Helsinki?

The Declaration of Helsinki expanded ethical guidance for medical research and emphasized the need for independent ethical oversight.

What is the Belmont Report?

The Belmont Report was created partly in response to Tuskegee and emphasized core ethical principles such as respect for persons, beneficence or welfare, and justice.

What is the TCPS 2?

The Tri-Council Policy Statement is the main Canadian ethical policy for research involving humans. It is based on respect for persons, concern for welfare, and justice.

Which agencies are connected to the TCPS?

The TCPS applies to research funded by SSHRC, CIHR, and NSERC, and institutions must have Research Ethics Boards to review protocols.

What does APA Standard 8.01 require?

It requires researchers to obtain institutional approval before beginning research.

What does APA Standard 8.02 require?

It requires informed consent, including information about the purpose of the study, risks, and the right to withdraw.

What does APA Standard 8.03 address?

It addresses consent for recording participants’ voices or images, unless an exception applies.

Under what circumstances can informed consent be waived?

Under APA 8.05, informed consent may sometimes be waived in minimal-risk research when specific ethical criteria are met.

When is deception allowed in research?

Under APA 8.07, deception is allowed only when it is justified by significant scientific value and when nondeceptive alternatives are not feasible.

When is correlational research especially useful?

It is useful when manipulation is impossible, impractical, or unethical, or when the researcher is interested in describing relationships rather than establishing causation.

Why can correlational research not establish causation?

Because the variables are only measured, not manipulated, so directionality and third-variable problems remain unresolved.

Why is it important to describe single variables even when studying relationships?

Because understanding each variable’s distribution helps researchers interpret the data accurately and detect issues such as skew, outliers, or restricted range.

What is a frequency table?

A frequency table lists the levels or scores of a variable and shows how often each one occurs.

What is a grouped frequency table?

A grouped frequency table combines nearby values into ranges of equal width, which is useful when there are many possible scores.

What is a histogram?

A histogram is a graph showing the distribution of a quantitative variable, with scores on the x-axis and frequencies on the y-axis.

What does it mean for a distribution to be unimodal?

A unimodal distribution has one clear peak.

What does it mean for a distribution to be bimodal?

A bimodal distribution has two distinct peaks, suggesting two common clusters of scores.

What is a symmetrical distribution?

A symmetrical distribution has left and right sides that are roughly mirror images.

What is a positively skewed distribution?

A positively skewed distribution has most scores at the lower end and a long tail extending toward higher scores.

What is a negatively skewed distribution?

A negatively skewed distribution has most scores at the upper end and a long tail extending toward lower scores.

What is central tendency?

Central tendency refers to the point around which scores tend to cluster. It is often thought of as the “average.”

Why is the mean often preferred?

The mean usually gives a strong indication of central tendency and has useful mathematical properties for further statistical analysis.

When is the median especially useful?

The median is especially useful when a distribution is skewed or contains outliers, because it is less affected by extreme scores than the mean.

When do mean, median, and mode tend to be similar?

They tend to be similar in a symmetrical, unimodal distribution.

Why can the mean be misleading?

The mean is strongly influenced by outliers, so one extreme value can pull it away from where most scores actually cluster.

What is variability?

Variability describes how spread out scores are around the centre of a distribution.

What is the standard deviation?

The standard deviation is the average distance of scores from the mean. It is the most common measure of variability.

Why do software packages often divide by N-1 instead of N when calculating standard deviation?

Because dividing by N-1 corrects for bias when estimating a population standard deviation from a sample.

What is a correlation?

A correlation describes the strength and direction of the relationship between two variables.

What is Pearson’s r?

Pearson’s r is a statistic that measures the direction and strength of a linear relationship between two quantitative variables.

How are correlations commonly interpreted in size?

Correlations around ±.10 are considered small, around ±.30 medium, and around ±.50 large.

How are correlations typically visualized?

They are often visualized using scatterplots, where each point represents one participant’s scores on two variables.

What is the difference between content validity and criterion validity?

Content validity asks whether the measure covers the full construct, while criterion validity asks whether the measure relates to external outcomes or variables as expected.

What is the difference between convergent and discriminant validity?

Convergent validity means the measure correlates with similar constructs, while discriminant validity means it does not correlate too strongly with different constructs.

What is the difference between test-retest reliability and internal consistency?

Test-retest reliability is consistency across time, while internal consistency is consistency across items within the same measure.