Statistics in Psychology and Everyday Life Study Notes
Module 8: Statistical Reasoning in Everyday Life
Learning Targets
- 8-1 Explain why we need statistics in psychology and in everyday life.
- 8-2 Describe descriptive statistics.
- 8-3 Explain how we describe data using the three measures of central tendency.
- 8-4 Discuss the relative usefulness of the two measures of variation.
- 8-5 Describe inferential statistics.
- 8-6 Explain how we determine whether an observed difference can be generalized to other populations.
The Need for Statistics
8-1 Why do we need statistics in psychology and in everyday life?
- Role in Psychology:
- Psychologists use statistics in descriptive, correlational, experimental, and other research designs to measure variables and interpret results.
- Importance of Statistical Understanding:
- Accurate statistical understanding is crucial for everyone, enhancing clarity and critical thinking about data in daily life.
- Memorizing complicated formulas is not necessary to think critically about statistical data.
- Risks of Misinformation:
- Quick estimates can lead to misrepresentation of reality, creating public misinformation.
- Examples of misinformation include assumptions about the percentage of people who are gay and the usage of brain capacity:
- Big claims such as “10 percent of people are gay” compared to “3 to 4 percent” from national surveys.
- The idea that we only use “10 percent of our brain,” which is misleading as we utilize nearly all parts.
- Doubt Big Numbers:
- Question the validity of large, round numbers presented without documentation (e.g., claims about missing children or homeless individuals) as they might be motivated by bias.
- Goal-Setting Bias:
- People are more inclined to set goals with round numbers (e.g., losing 20 pounds rather than 19 or 21)
- Statistical behavior modification is nearly four times more effective when focusing on the .300 average rather than .299.
- Statistical Illiteracy and Health Risks:
- Misinformation can lead to health scares, exemplified by media reporting on contraceptive pills causing a supposed 100% increase in blood clot risks. The actual increased risk was from 1 in 7000 to 2 in 7000, highlighting the need for accurate statistical reasoning.
Descriptive Statistics
8-2 What are descriptive statistics?
- Definition:
- Descriptive statistics summarizes and describes characteristics of a group under study, similar to how teachers assess student performance.
- Data Visualization:
- Researchers represent data using visual formats like histograms, which display distributions (e.g. truck durability after 10 years).
- Graphical Representation:
- Caution: Read scale labels carefully, as misleading graph designs can exaggerate or minimize differences:
- Example of two graphs differing only in their y-axis scale, where a limited range can make differences appear more pronounced (Graph (a)) than they actually are (Graph (b)).
- Key Point: Always verify the scale labels when interpreting graphs to comprehend the data accurately.
Measures of Central Tendency
8-3 How do we describe data using the three measures of central tendency?
- Definition of Central Tendency:
- A single score representing a set of scores.
- Three main measures: Mode, Mean, and Median:
- Mode: The most frequently occurring score or scores.
- Mean: The arithmetic average, calculated by summing all scores and dividing by the number of scores.
- Median: The middle score in an ordered distribution; it divides the scores into two equal halves.
- Symmetrical Distributions:
- In a symmetrical bell-shaped distribution, the mode, median, and mean are often similar.
- Skewed Distributions:
- In skewed distributions, the mean can be disproportionately affected by extreme values, leading to significant discrepancies:
- Example with income data wherein high incomes disproportionately raise the mean compared to the median.
- Important Consideration:
- Recognize that the mean can misrepresent actual data trends, especially in populations with significant outliers.
Measures of Variation
8-4 What is the relative usefulness of the two measures of variation?
- Understanding Variation:
- Averages can be misleading if not accompanied by measures of variation, which indicate how scores differ from one another.
- Importance of Low Variability:
- Averages based on scores with low variability provide more reliable predictions (e.g., a consistent scoring basketball player).
- Range as a Measure of Variation:
- The range (difference between highest and lowest scores) is often imprecise as it can overlook variations in other scores.
- Standard Deviation:
- A more useful measure that examines how much individual scores differ from the mean:
- It takes all scores into account.
- Example illustrating differences in standard deviation between classes, where one has clustered scores (low variability) and the other has widely dispersed scores (high variability).
Inferential Statistics
8-5 What are inferential statistics?
- Definition and Purpose:
- Inferential statistics enables researchers to infer conclusions about a population based on sample data, evaluating the reliability of observed differences.
- Understanding Noise in Data:
- Averages may differ between groups not due to genuine differences but chance fluctuations.
- Reliability of Observed Differences:
- Inferential statistics help assess how likely it is that observed differences reflect real differences rather than random chance.
When Is an Observed Difference Reliable?
8-6 How do we know whether an observed difference can be generalized to other populations?
- Principles for Generalization:
- Representative Samples:
- Effective generalizations stem from samples that accurately represent the relevant population rather than exceptional or biased samples.
- Variability:
- Observations with low variability yield more reliable averages than those with high variability (examples illustrate this point).
- Sample Size:
- Larger samples tend to be more reliable than smaller ones.
- Caution Against Anecdotes:
- Generalizations from anecdotal evidence or small samples can lead to erroneous conclusions.
- Statistical Significance and Reliability:
- Statistically significant findings indicate observable differences are likely not due to chance; however, they must also be practical in importance.
- Illustrative Examples:
- Large sample studies that yield statistically significant results but are minor in real-world impact (e.g., social media behavior) must be evaluated for practical relevance.
- Conclusion:
- Smart analytical thinking requires distinguishing between statistical significance (likelihood of being due to chance) and practical significance (real-world relevance) in research findings.