Question

what are weaknesses to only describing data at the level of measurement and using graphs?

Answer 1

Only Describing data at levels of measurement(nominal/ordinal/interval/ratio) or only using graphs are useful to have a preliminary idea of how the data behaves but does not provide enough information. Most graphing and charting methods can represent only discrete observations such as bar graphs which would be inappropriate for data points with decimals. Graphs give a general visual idea of the data, but fails to give actual measures of metrics.

In using level of measurement, the drawbacks can be misspecification of units, two groups can be compared only if their units are same as in Interval and Ratio Scale. Further, let us take nominal scale- this can only tell us about the mode of the data as it classifies the observations to groups. In Ordinal scale , the gaps between two observations in consecutive order may not be equal , as a result, comparison is impossible.