Handling data — Eduqas A-Level Psychology

Handling data covers the essential mathematical and graphical competencies required across the Eduqas Psychology specification. It encompasses identifying levels of measurement (nominal, ordinal, interval, ratio), calculating and choosing measures of central tendency and dispersion, and accurately plotting and interpreting graphs.

This topic forms the quantitative bedrock of Component 2 (Psychological Investigations) and Component 1 research methodology. Developing precision in data transformation, standard deviation analysis, and distribution identification ensures students can critically evaluate psychological evidence and secure maximum mathematical marks.

• →Levels of Measurement: Nominal (categorical), Ordinal (ordered/ranked without equal intervals), Interval (standardised, equal units without absolute zero), and Ratio (equal intervals with a true absolute zero).
• →Measures of Central Tendency: Mean (incorporates all values but sensitive to outliers), Median (middle value, robust against skewness), and Mode (most frequent category/value).
• →Measures of Dispersion: Range (difference between highest and lowest, sensitive to outliers) and Standard Deviation (calculates the spread of data points around the mean).
• →Distributions: Symmetrical bell curves (normal distribution where mean = median = mode) versus asymmetrical skewed distributions (positive tail right, negative tail left).
• →Graphical Presentation: Appropriate selection and accurate labelling of bar charts, histograms, scattergrams, and line graphs according to data type.

• Reports a correlation coefficient to two or three significant figures, for example r = 0.78 or r = 0.783.
• Counts significant figures from the first non-zero digit rather than from the decimal point.
• Rounds the final retained digit correctly by inspecting the next digit.
• Retains the sign of a negative coefficient when rounding, for example −0.46.
• Distinguishes significant figures from decimal places when justifying the reported value.
• Sums the raw scores within each condition before dividing.
• Divides each condition total by the number of participants in that condition.
• Keeps the two conditions separate and labels each calculated mean.
• Reports each mean with the same units as the raw data.
• Checks the arithmetic by re-adding the scores or multiplying the mean by the number of participants.
• Selects a bar chart for categorical frequency data because the categories are discrete and the bars are separated by gaps.
• Selects a histogram for continuous data grouped into equal class intervals because adjacent bars touch along a continuous scale.
• Constructs a frequency table that pairs each category or class interval with a tally and a frequency.
• Interprets a display by reading frequencies, identifying the modal category or interval and describing the overall shape.
• Justifies a chosen display by linking it to the nature of the data rather than to personal preference.
• Defines probability as a value between 0 and 1 expressing the chance of an event, with 0.05 equal to a 5% chance.
• States that the 0.05 level accepts a result as significant when the probability of it occurring by chance is 0.05 or less.
• States that the 0.01 level is stricter, requiring a probability of 0.01 or less before a result is treated as significant.
• Explains that the 0.01 level reduces the risk of a Type I error compared with the 0.05 level.
• Explains that the stricter 0.01 level increases the risk of a Type II error, so fewer genuine effects are detected.
• Identifies that a random sample gives every member of the target population an equal chance of being selected
• Describes assigning a unique number to each member of the target population and selecting numbers using a random number generator or random number table
• Explains that a stratified sample divides the target population into relevant subgroups and then samples randomly within each subgroup
• States that the number drawn from each stratum is proportional to that stratum's size in the target population
• Recognises that both methods aim to reduce bias and improve generalisability to the target population
• Acknowledges a limitation, such as stratified sampling depending on accurate information about the strata or neither method guaranteeing representativeness
• Defines the mean as the total of all scores divided by the number of scores
• Defines the median as the middle value when scores are ordered, or the mean of the two middle values for an even number of scores
• Defines the mode as the most frequently occurring value in a data set
• Explains that the mean uses all scores but is affected by extreme values, whereas the median is less affected by outliers
• Explains that the mode is appropriate for nominal or categorical data where the mean and median cannot be calculated
• Selects an appropriate measure for a given data set by referring to level of measurement, outliers or distribution shape
• Plots each participant's pair of scores as one point on a correctly labelled scatter diagram
• Identifies a rising pattern from bottom-left to top-right as a positive correlation
• Identifies a falling pattern from top-left to bottom-right as a negative correlation
• Identifies a shapeless spread of points with no trend as no correlation
• Labels both axes with the variable names and appropriate units
• States a null hypothesis and an alternative hypothesis that identify both conditions of the class experiment
• Selects a non-parametric test of differences appropriate to the related design and the level of measurement
• Ranks the differences between paired scores and sums the ranks for each direction separately
• Compares the calculated value with the critical value at the stated significance level and the correct number of pairs
• Accepts or rejects the null hypothesis and interprets the outcome in terms of the experiment
• Rounds the total score and the number of participants to convenient values before dividing.
• Divides the rounded total score by the rounded number of participants to obtain an approximate mean.
• States the estimated mean to an appropriate level of accuracy, such as one significant figure.
• Checks that the estimated mean falls within the possible range of test scores for the measure used.
• Identifies nominal data as named categories with no inherent order between them.
• Identifies ordinal data as ranked categories where the intervals between ranks are not equal.
• Identifies interval data as numerical scores with equal intervals between adjacent values.
• States the level of measurement used in a described study by referring to how the variable was recorded.
• Places the mean, median or mode at the central peak of a symmetrical bell-shaped curve.
• Identifies that in a normal distribution the mean, median and mode coincide at the same point.
• Recognises symmetry as the defining feature that separates a normal from a skewed distribution.
• States that a positive skew has a long tail towards high scores with the mean above the median and mode.
• States that a negative skew has a long tail towards low scores with the mean below the median and mode.
• Names an inferential test that matches the level of measurement, the design and the question asked (difference or correlation) in the described investigation
• States the level of measurement of the data and links it explicitly to the choice of test
• States whether the design is independent measures or repeated measures/matched pairs and links it explicitly to the choice of test
• Explains why at least one alternative test would not be appropriate for these data
• Reports the calculated value alongside the critical value at the stated significance level and degrees of freedom
• Interprets the comparison as significant or non-significant and relates the outcome to the experimental or alternative hypothesis
• Identifies the statistical test in the question and applies the correct direction of comparison for that test
• Locates the critical value using the stated N or degrees of freedom and the 0.05 column for a one-tailed test
• Compares the observed value with the critical value and states whether the result is significant
• Concludes that the null hypothesis is rejected when the result is significant and retained when it is not
• Defines the range as the difference between the highest and lowest score in a set
• Defines the standard deviation as a measure of the average spread of scores around the mean
• Explains that the range depends only on the two extreme values and is therefore sensitive to outliers
• Explains that the standard deviation uses every score, so it is less distorted by a single outlying value
• Applies the comparison to a specific set of scores containing an outlier to justify why the standard deviation is more useful
• Acknowledges a limitation of the standard deviation, such as the greater effort needed to calculate it
• Defines qualitative data as non-numerical material such as words, descriptions or interview transcripts that capture meaning and experience
• Defines quantitative data as numerical material such as counts, scores or measurements that can be analysed statistically
• Contrasts the two forms in terms of richness of detail against ease of comparison and statistical analysis
• Explains that a transcript can be converted by defining categories or themes and coding each response into them
• Describes counting the frequency of a coded theme or rating responses on a numerical scale to produce scores
• Acknowledges that conversion loses some qualitative richness while gaining measurability and comparability
• Defines primary data as data collected first-hand by the researcher for the current study
• Defines secondary data as existing data gathered by others for a different purpose and reused
• Identifies that a researcher dealing directly with participants collects primary data
• Distinguishes the two by who collected the data and the purpose of collection

• 💡Underline the first non-zero digit before counting significant figures.
• 💡Circle the digit after your last retained figure to decide whether to round up.
• 💡State the number of significant figures you have used so the marker can see your choice.
• 💡Write out the scores for each condition in a column before adding.
• 💡Show the sum and the divisor so your working is visible.
• 💡Multiply your mean by the number of participants to check it returns the total.
• 💡Name the data type first (categorical or continuous) before choosing a display, and write that reason in your answer.
• 💡Check that the frequencies in your table add up to the total number of scores before sketching anything.
• 💡When interpreting, quote the modal category or interval and one other feature such as skew or range of frequencies.
• 💡Convert each level to a percentage (0.05 to 5%, 0.01 to 1%) to make the comparison concrete.
• 💡Link each level to the matching error type: 0.05 raises Type I risk, 0.01 raises Type II risk.
• 💡Use the phrase 'probability of occurring by chance' to keep the explanation precise.
• 💡Define the target population explicitly before describing the sampling procedure
• 💡Name the randomising device, such as a random number generator or random number table, rather than saying 'chosen at random'
• 💡For stratified sampling, state the strata, the proportional allocation and the random selection within each stratum
• 💡Distinguish clearly between the sampling method and its strengths or limitations
• 💡State the level of measurement of the data before choosing a measure of central tendency
• 💡Justify your choice by referring to outliers, skew or the type of data, not just by preference
• 💡Show the ordered list when finding a median so the middle value is clear
• 💡Check whether a data set has more than one mode or no mode at all
• 💡State the direction of the pattern first, then justify it by referring to how the points run across the diagram
• 💡Check that both axes carry a variable name and units before deciding on the correlation
• 💡Distinguish clearly between no correlation and a weak correlation by describing the spread of the points
• 💡Write the hypotheses before any calculation so the direction of the test is fixed
• 💡Show the ranking of differences and the two rank sums so the calculated value can be followed
• 💡Finish with a sentence that links the statistical decision back to the class experiment
• 💡Show the rounded values you used so the approximation is visible to the examiner.
• 💡Perform the division mentally with rounded numbers rather than reaching for a calculator, since the point is a quick estimate.
• 💡Sanity-check the answer against the range of possible scores before selecting an option.
• 💡Ask whether the values are names, ranks or measured quantities with equal units before choosing a level.
• 💡Justify the chosen level by pointing to a feature of the variable, such as equal intervals or absence of order.
• 💡Eliminate options that confuse ordinal with interval by checking whether the gaps between values are equal.
• 💡Check symmetry before placing any measure: a symmetrical curve means all three averages share one point.
• 💡For a skewed curve, decide the tail direction first, then order mean, median and mode accordingly.
• 💡Read the question carefully to see whether it asks for the mean, median or mode before marking.
• 💡Sketch the curve lightly so the axis of symmetry is visible when you place your answer.
• 💡Write the justification as a short chain: measurement level, then design, then difference or correlation, then the named test
• 💡Use the correct technical name of the test and spell it accurately
• 💡State the significance level and degrees of freedom you used before comparing values
• 💡Write the observed value and the critical value side by side before making the comparison, so the direction is explicit
• 💡State the decision in full: whether the result is significant and whether the null hypothesis is rejected or retained
• 💡Check the table's row and column headings carefully, since N and degrees of freedom are not interchangeable
• 💡Quote the significance level and tailedness you used so the examiner can see the comparison is correctly set up
• 💡Define both measures precisely before comparing them, so the comparison rests on accurate terms
• 💡Anchor the explanation in the specific scores provided, naming the outlier and showing how it affects the range
• 💡Balance the answer by noting one limitation of the standard deviation as well as its advantage
• 💡Define both terms precisely before contrasting them, so the difference is explicit
• 💡Use a concrete example such as coding an interview transcript for mentions of a named theme
• 💡Explain the coding step by step: define categories, code responses, then count or score them
• 💡State one strength and one limitation of converting qualitative data to quantitative data
• 💡Read the scenario carefully and ask who collected the data and for what purpose
• 💡Apply the rule that direct contact between researcher and participants yields primary data
• 💡Eliminate options that describe reused existing material when the scenario shows first-hand collection
• 💡Always clearly label graph axes including units of measurement (e.g. 'Mean recall score (out of 20)') and provide an informative, contextual title.
• 💡Show all stages of your mathematical calculations; examiners award method marks even if a minor arithmetic error occurs in the final answer.
• 💡Whenever asked to compare conditions, comment on both central tendency (average performance) and dispersion (consistency of scores) with comparative language.

• Counts decimal places instead of significant figures; count from the first non-zero digit.
• Rounds 0.7834 to 0.8 and calls it two significant figures; keep two significant figures as 0.78.
• Drops the minus sign when rounding a negative coefficient; retain the sign in the reported value.
• Reports a coefficient to five or six significant figures; reduce it to two or three significant figures.
• Divides by the total number of participants across both conditions; divide by the number in each condition.
• Omits one raw score when adding; list every score before summing.
• Mixes scores from the two conditions; keep each condition's data separate.
• Reports a mean without units or condition label; attach both to the value.
• Draws a histogram for named categories such as attachment types; use a bar chart with gaps because the categories are discrete.
• Leaves gaps between the bars of a histogram; make adjacent bars touch because the underlying scale is continuous.
• Uses unequal class intervals without adjusting bar width; keep intervals equal or scale the width to the interval.
• Reads a bar height as a raw score rather than a frequency; state that the height shows how many scores fall in that category.
• Treats 0.05 as a probability of 50%; read it as 5 in 100, that is, a 5% chance.
• Claims the 0.01 level makes a Type I error more likely; state that the stricter level makes a Type I error less likely.
• Says a result below 0.05 proves the hypothesis is true; describe it as unlikely to have arisen by chance, not proven.
• Confuses the significance level with the observed probability; keep the pre-set threshold separate from the calculated value.
• Confuses random sampling with haphazard or opportunity sampling; define the population and use a genuine random selection device
• Describes a stratified sample without random selection within strata; include random selection inside each subgroup
• Omits proportionality when describing stratified sampling; state that each stratum contributes in proportion to its size
• Claims a random sample is always representative; note that it reduces bias but does not guarantee representativeness
• Calculates the median without first ordering the scores; arrange the values in ascending order before finding the middle
• Uses the mean for nominal data such as categories of favourite subject; select the mode instead
• Assumes the mean is always best because it uses all scores; consider outliers and level of measurement
• Forgets to average the two middle values for an even number of scores; take the mean of those two values
• Joins the points with a line as if the diagram were a line graph; leave the points unconnected and read the overall pattern
• Places each variable's scores on separate axes without pairing them per participant; plot one point per participant using both scores together
• Describes a single extreme point as the correlation; describe the trend shown by the majority of points
• Chooses a test of correlation for a design that compares two conditions; select a test of differences when the design compares conditions or groups
• Omits the significance level and the number of pairs when finding the critical value; state both before comparing values
• Reports the calculated value without deciding about the hypotheses; state whether the null hypothesis is accepted or rejected and say what this means
• Multiplies the total score by the number of participants instead of dividing; divide the total by the number of participants to estimate the mean.
• Reports an over-precise figure such as 59.939 from rounded inputs; state the estimate to a sensible accuracy such as one significant figure.
• Ignores whether the estimate is plausible for the scoring range; compare the result with the minimum and maximum possible scores.
• Treats ranked categories as interval data; recognise that unequal gaps between ranks make the data ordinal.
• Classifies named groups such as gender or attachment type as ordinal; recognise that categories without order are nominal.
• Describes a level of measurement without linking it to the variable in the study; state the level and justify it from how the scores were recorded.
• Places the mean at one tail of a normal curve; correct this by marking the central axis of symmetry where the peak lies.
• Assumes the mean always equals the median in any distribution; restrict this to symmetrical normal distributions.
• Reverses the direction of skew; remember a positive skew tails towards high scores and a negative skew tails towards low scores.
• Confuses the mode with the highest frequency value's position on a skewed curve; check which measure the long tail pulls furthest.
• Chooses a test because it is familiar rather than because it matches the level of measurement and design; justify the choice from the features of the data
• Treats whether the hypothesis is one-tailed or two-tailed as the factor that fixes which test is chosen; use tailedness only when interpreting the critical value, and select the test from measurement level, design and difference-versus-correlation
• Applies a greater-than-or-equal rule to the sign test, Wilcoxon or Mann-Whitney U; use less than or equal to the critical value for these tests
• Applies a less-than-or-equal rule to chi-squared or Spearman's rho; use greater than or equal to the critical value for these tests
• Reads the two-tailed column when the question specifies a one-tailed test; select the column matching the stated tailedness and significance level
• Uses the wrong row by confusing N with degrees of freedom; check which quantity the table requires for that test
• Describes the range as a measure of central tendency; classify it as a measure of dispersion instead
• Claims the standard deviation ignores outliers entirely; explain that it is less affected because it uses all scores, not that it excludes them
• Gives definitions without applying them to the given scores; link the comparison to the specific data set and its outlier
• Treats the standard deviation as always preferable; note its greater calculation effort as a trade-off
• Treats qualitative and quantitative data as identical because both come from participants; distinguish the form of the data, words against numbers
• Assumes conversion is impossible or invalid; explain that coding and counting themes produces usable numerical data
• Describes conversion without naming a method; specify coding into categories and counting or rating responses
• Claims quantitative data always gives deeper insight; note that numbers gain comparability but lose qualitative detail
• Labels data from direct contact with participants as secondary; recognise that first-hand collection makes it primary
• Confuses the source of data with its type; judge primary or secondary by who collected it and why
• Assumes all published data is primary; note that reused existing data is secondary
• Confusing bar charts with histograms: Bar charts represent discrete categorical data and must feature gaps between bars, whereas histograms represent continuous data with no gaps.
• Believing the mean is always the best measure of central tendency: While it uses all data, it is heavily skewed by extreme outliers, making the median preferable when data is non-normal or skewed.
• Assuming range includes adding one: In Eduqas psychological contexts, standard mathematical subtraction (highest minus lowest) is typically accepted, but candidates must show workings clearly.

1. 1Day 1-2: Review levels of measurement (NOIR) with flashcards and classify 15 different psychological variables.
2. 2Day 3-4: Practice calculating mean, median, mode, and range from raw data sets, focusing on justifying which measure fits specific data sets.
3. 3Day 5-6: Master standard deviation interpretation and distribution curve sketches (normal, positive, negative skews).
4. 4Day 7: Complete timed past-paper data handling questions from Eduqas Component 2 papers, checking against official mark schemes.

• 📋Calculation questions: 2-4 mark questions requiring candidates to calculate percentages, central tendencies, or ratios from given stimulus data.
• 📋Graphical representation questions: 3-4 mark questions requiring candidates to sketch or accurately plot a graph based on provided summary tables.
• 📋Data interpretation and evaluation: 4-6 mark questions asking students to draw conclusions about experimental conditions using both mean and standard deviation values.