Research Methods - Analysing Research — OCR GCSE Psychology

Analysing research covers the mathematical and interpretive techniques psychologists use to make sense of empirical data gathered from experiments, self-reports, observations, and correlations. This includes calculating descriptive statistics such as measures of central tendency and dispersion, as well as converting raw figures into fractions, percentages, and ratios.

A thorough understanding of data analysis is vital across the OCR GCSE Psychology specification because research methods questions account for a significant proportion of marks in both Paper 1 and Paper 2. Being able to evaluate strengths and weaknesses of different analytical tools ensures students can critically assess scientific claims in both exam questions and real-world psychological studies.

• →Measures of Central Tendency: The mean (arithmetic average), median (middle score in an ordered set), and mode (most frequent score) each provide distinct insights and have specific limitations regarding outlier sensitivity.
• →Measures of Dispersion: The range assesses data spread by calculating the difference between the highest and lowest values, highlighting variability across conditions.
• →Data Representation: Bar charts display discrete categories, histograms represent continuous grouped data, and scatter diagrams illustrate the strength and direction of relationships between co-variables.
• →Qualitative vs. Quantitative Data: Quantitative data is numerical and directly measurable, whereas qualitative data is descriptive, rich in detail, and typically analysed using thematic analysis.

• Identifies quantitative data as numerical information recording amounts or scores, such as a memory test score.
• Identifies qualitative data as non-numerical information recording qualities or meanings, such as a written description.
• Distinguishes primary data as data collected first-hand by the researcher for the current study.
• Distinguishes secondary data as existing data reused by the researcher, such as published statistics.
• States a strength of quantitative data, such as ease of comparison and statistical analysis.
• States a strength of qualitative data, such as rich detail and insight into participants' meanings.
• States a strength of primary data, such as being collected for the study's exact aim with controlled quality.
• States a strength of secondary data, such as saving time and cost and offering large samples.
• Calculates the mean by summing all values and dividing by the number of values.
• Finds the median by ordering the values and taking the middle one, averaging the two middle values when the count is even.
• Identifies the mode as the most frequent value and the modal class as the group with the highest frequency.
• Calculates the range as the highest value minus the lowest value.
• Expresses a comparison as a ratio, percentage or fraction, such as 3:2 or 60%.
• Converts a value into standard form, such as writing 4500 as 4.5 × 10³.
• Rounds an answer to a stated number of decimal places or significant figures.
• Describes a normal distribution as symmetrical and bell-shaped with most scores near the mean.
• Uses a summary statistic to estimate a typical value from collected data.
• Selects a bar chart for frequencies of separate categories and a histogram for a continuous variable with touching bars
• Reads a pie chart by relating each sector angle to its share of the total of 360°
• Identifies a scatter diagram as the display for paired scores on two variables
• Uses a frequency table with tallies to record how often each value or category occurs
• Distinguishes a line graph, which joins points across a continuous scale, from a bar chart of categories
• Defines internal reliability as the consistency of items within a single test or scale
• Defines external reliability as consistency of a measure across occasions or researchers
• Defines inter-rater reliability as agreement between two or more observers scoring the same behaviour
• Defines ecological validity as the extent to which findings generalise to real-life settings
• Defines population validity as the extent to which findings generalise from the sample to the target population
• Defines construct validity as the extent to which a measure reflects the theoretical construct it claims to measure
• Defines demand characteristics as cues that lead participants to guess the aim and alter behaviour
• Defines the observer effect as change in behaviour caused by awareness of being observed
• Defines social desirability as the tendency to present oneself in a favourable light
• Identifies gender bias as drawing conclusions about one sex from data collected mainly from the other
• Identifies cultural bias as using tasks, norms or interpretations that favour one cultural group
• Identifies age bias as generalising findings from one age group to another
• Identifies experimenter bias as the researcher's expectations influencing participants or the recording of results
• Identifies observer bias as the recorder interpreting behaviour to match expectations
• Identifies bias in questioning as leading, loaded or ambiguous questions steering responses
• Explains how the named bias distorts the findings rather than merely naming it

• 💡Name the axis you are using (numerical versus non-numerical, or first-hand versus existing) before giving the example.
• 💡Give a concrete example with each type, such as 'reaction time in milliseconds' for quantitative data.
• 💡When asked for a strength, link it to a purpose, such as ease of comparison for quantitative data.
• 💡Check whether a scenario describes who collected the data before deciding primary or secondary.
• 💡Show your working for the mean so a single arithmetic slip does not lose all the credit.
• 💡Order the data explicitly before locating the median, especially with an even number of values.
• 💡Check the required rounding, then state the answer to that many decimal places or significant figures.
• 💡Sketch the normal curve to justify a statement about where most scores lie.
• 💡Name the data type first (category, continuous or paired) and let that decide the display
• 💡Check whether bars touch or have gaps, because that detail separates histograms from bar charts
• 💡For pie charts, confirm the sectors add to a full circle before reading any single share
• 💡Pair each type of reliability with its assessment method: split-half for internal, test-retest for external, correlation between raters for inter-rater
• 💡Pair each type of validity with the question it answers: real world, other populations, correct construct
• 💡Distinguish the three participant effects by their cause: cues about aim, awareness of observer, wish to look good
• 💡Use a short scenario and name which reliability, validity or effect applies
• 💡Match each scenario to the specific named bias before choosing an option, rather than picking the first plausible word
• 💡Ask who is affected and how the distortion enters the study, then check the option against that reasoning
• 💡Eliminate options that describe sampling size or ethics when the question asks specifically about bias
• 💡Always include correct units when calculating descriptive statistics (e.g. 'seconds', 'errors', or '%') to prevent unnecessary loss of precision marks.
• 💡When asked to justify a choice of graph, mention both the type of data collected (e.g. discrete vs continuous) and what the graph illustrates (e.g. comparing group differences vs plotting a correlation).

• Confuses primary data with quantitative data; keep the two axes separate, since primary data can be either numbers or words.
• Treats secondary data as automatically unreliable; recognise that secondary sources vary in quality and can be strong when reputable.
• Assumes qualitative data cannot be analysed; note that qualitative data can be analysed thematically even without statistics.
• Labels a rating scale as qualitative because it appears on a questionnaire; classify it as quantitative because it yields numbers.
• Forgets to order the values before finding the median; sort the data first, then take the middle value.
• Divides by the wrong count when calculating the mean; divide by the number of values, not the highest value.
• Confuses the modal class with the mode of raw scores; for grouped data report the class with the highest frequency.
• Writes standard form with the wrong exponent; check that 4500 becomes 4.5 × 10³, not 4.5 × 10².
• Chooses a histogram for named categories; use a bar chart with gaps for categories and reserve histograms for continuous data
• Treats pie chart angles as raw frequencies; convert each frequency into a share of the 360° total
• Plots paired scores on a line graph; use a scatter diagram when two variables are measured on the same participants
• Describes internal reliability as consistency across repeated testing occasions; keep internal reliability to consistency among items within one test
• Confuses inter-rater reliability with inter-item consistency; keep inter-rater reliability to agreement between separate observers
• Treats ecological validity as realism of the task alone; link it to generalisation to real-life settings
• Confuses demand characteristics with the observer effect; keep demand characteristics to cues about the aim and the observer effect to awareness of being watched
• Confuses observer bias with experimenter bias; keep them distinct by noting that observer bias concerns interpretation while recording and experimenter bias concerns the researcher's influence on participants or data
• Treats a small sample as automatically a source of bias; recognise that the listed sources are specific and a small sample is a sampling issue unless it produces one of the named biases
• Assumes bias in questioning means participants lie; recognise that it concerns how questions are worded and framed
• Names a bias without saying what it distorts; add a sentence explaining the direction of the distortion
• Students often believe the mean is always the 'best' measure of central tendency; however, extreme anomalous scores skew the mean, making the median a more representative measure of typical performance in skewed data.
• A common error is treating bar charts and histograms identically; bar charts must have gaps between bars to reflect discrete categories, whereas histogram bars touch to represent continuous intervals.

1. 1Day 1-2: Review and practice calculations for mean, median, mode, and range using varied datasets including even numbers and outliers.
2. 2Day 3-4: Practice drawing and interpreting bar charts, frequency tables, and scatter diagrams, paying special attention to accurate axis labelling and scale.
3. 3Day 5: Compare the strengths and weaknesses of each measure of central tendency and dispersion in relation to specific research scenarios.
4. 4Day 6-7: Complete timed OCR past-paper exam questions focused on research analysis and calculate percentages, ratios, and fractions from raw study tables.

• 📋Calculation questions: Requiring students to compute the mean, median, mode, range, ratio, or percentage from raw data sets (typically 1-3 marks).
• 📋Data interpretation and evaluation: Asking students to explain why a researcher chose a specific measure of central tendency or graph, referencing the study's data characteristics (typically 2-4 marks).
• 📋Graph plotting and reading: Plotting specific data points on pre-drawn axes or describing the pattern shown by a bar chart or scatter plot (typically 2-4 marks).