Data Analysis
The undertaken data analyses the experimental outcomes with varying audio interface of the footwear and observing the changes in the behaviour of users. In this assignment, the researcher analyses the data set gathered from a footwear user study considering the physical aspects of gender, weight, shoe size and height. The participants received the high frequency, low frequency and controlled audio feedback from walking at the time of wearing the prototype shoes. For each cases of this experiment, the researcher captured the perceptions of the participants about their body weight, mood, emotions and changes in behaviour involving three dimensions. The commonly used models that are valence, arousal and dominance.
The software MS-Excel is used for analysing the data set.
The data cleaning process is very crucial and necessary aspect to data analysis. Before, using the data for analysis purpose, an analyst should focus on data cleaning for having correct information. As per data sheet, the data set has many variables such as Gender, Age, “ShoeSizeUK”, “Weight_Kg”, “Height_cm”, “GSR_Z-scores”, “BodyVisualization_LOGscore”, “HeelPressure_Zscore”, “ToePressure_Zscore”, “FootAcceleration_Zscore”, “Valence”score, “Arousal” score, “Dominance” score and questionnaires regarding several variables alike speed, weight, strength, straightness, agency, vividness, surprise and “FeetLocalization”. It is observed that-
- Sample 4 has lots of missing values starting from L5 to AC5.
- Sample 2 and 8 has missing values in the range AD3 to AF2.
- Besides, sample 16, 11 and 10 also have some missing values.
The missing data appeared as measurements of replication was carried out for some of the samples completely. For single empty response of any sample, the analyst manually put the mean value of the column in the blank space. For the sample whose number of missing values are more than 3, the analyst, has decided to omit the whole sample from the data set. For example, sample 4 is deleted from the entire data set but the missing value of “GSR_Zscore” with high frequency of sample 2 is replaced by the average value of the entire AD column (Hand 2007). The outcomes with the presence of missing value may be misleading. The quantitative data is therefore removed.
To have descriptive statistics, the study consists with 21 sample that have 17 females and 4 males. The response of these people is collected as per survey questionnaires. The survey has different factors such as Age, Size of shoes of the responders, Weights of responders in Kg and Heights of responders in cm.
The average age of the responders is found to be 24.36 years (SD = 4.86 years). The middle most value of age is 22.5. The age ranges from 18 to 35 years. It is 95% evident that average Age lies in the interval of 26.52 years to 22.21 years. The mean shoesize of the sampled responders is 6.023 unit (SD = 1.829 unit). The 50% of the values lie above the shoe size 5.75 units. The shoe size ranges in the interval of 3 units to 10 units. The estimated average shoe size of all the samples lie in the interval of 6.83 units to 5.21 units. The weight of the sampled people is 59.25 Kgs (SD = 10.546 Kgs). The middle most value of the weight of the responders is 57 Kgs. The maximum weight of any responder was found to be 89 Kgs and minimum weight of any responder was found to be 47 Kgs. The mean height of all the sampled persons lie in the interval of 54.574 Kgs and 63.926 Kgs. The heights of the sampled responders have average 164.818 cm (SD = 6.905 cm). The median and modal value of the heights of the responders are both 165 cm. The estimated mean with 95% probability lies in the interval of 161.757 cm and 167.880 cm. The “GSR_Zscore” is calculated with the average of “GSR_Zscore” with high frequency, low frequency and control. The average GSR or Galvanic skin response is (-0.01251) (SD = 0.327119). The lower and upper control limit of estimated in the range (-0.16141 and 0.136393) with 95% probability.
Descriptive Statistics
The data analysis table as per “Galvanic Skin Response” discovers that z-score of Galvanic Skin Response is greater in case of High Frequency and lowest in Low Frequency. The margin of error is highest for control group and least for high frequency group. More of it is discovered that-
- The average z-score of the high frequency Galvanic Skin Response lies in the interval of 0.21 to 0.04 with 95% probability (Morey 2008).
- The average z-score of the low frequency Galvanic Skin Response lies in the interval of 0.017 to (-0.29) with 95% probability.
- The average z-score of the Galvanic Skin Response of control lies in the interval of 0.26 to (-0.31) with 95% probability.
The bar chart of Galvanic Skin response is overall highest in case of high frequency and lowest in case of low frequency. However, the range of z-score is highest in case of Galvanic Skin Response in control state.
The positive Valence response is higher in case of the responses that have response rate greater than 5. Out of 21 samples, 17 samples have higher valence rate, 11 samples have low valence rate and 13 samples have positive rate of response. The proportions of Valence response in high frequency, low frequency and controlled frequency are 0.81, 0.52 and 0.62 respectively.
- With 95% possibility, the proportion of positive response ranges in the interval of 0.64 to 0.98 with high frequency.
- With 95% possibility, the proportion of positive response ranges in the interval of 0.74 to 0.31 with low frequency.
- With 95% possibility, the proportion of positive response ranges in the interval of 0.83 to 0.41 with controlled frequency.
The positive perceived feet localization are the samples that have feet localization measure more than 4 from two replications. The proportion of positive feet localization is greater for High audio frequencies and lower for Controlled audio frequencies. The estimated feet localization with high audio frequency varies from the proportion 0.78 and 1. The estimated feet localization with control audio frequency ranges within the proportion 0.52 to 0.91.
The bar chart clearly indicates that the proportion of positive feet localization is greater for high audio frequency followed by low audio frequency. The proportion of positive feet localization is lowest for controlled audio frequency.
It could be concluded that perceived positive feet localization varies with different audio frequencies significantly.
The positive perceived Vividness are the samples that have Vividness measure more than 4. The proportion of positive surprise feeling is greater for Low audio frequencies and lower for both High and Controlled audio frequencies. The estimated Vividness feeling with low audio frequency varies from the proportion 0.53 and 0.13. The estimated Vividness response with both high and control audio frequency that ranges within the proportion 0.48 to 0.09.
The bar chart firmly refers that the proportion of positive surprise feeling is greater for high audio frequency followed by low audio frequency. The proportion of positive surprise feeling is lowest for controlled audio frequency.
Therefore, perceived positive vividness response change with various audio frequencies insignificantly.
5.3. Research Question: Do the Z-scores of the three variables “Heel Pressure”, “Toe Pressure” and “Foot Acceleration” equally respond according to the three types of audio frequency that are High, Low and Control?
The averages of Z-score of two replications for Heel pressure refers that-
- The average Z-score is higher for Low audio frequency followed by Controlled audio frequency. The average Z-score is least for high frequency.
The averages of Z-score of two replications for Toe pressure indicates that-
- The average Z-score is higher for Low audio frequency followed by High audio frequency. The average Z-score is least for controlled audio frequency.
The averages of Z-score of two replications for Foot acceleration refers that-
- The average Z-score is greater for high audio frequency followed by controlled audio frequency. The average Z-score is therefore least for Low audio frequency.
Conclusion:
Hence, it can be concluded that the Z-scores of the three considered variables “Heel Pressure”, “Toe Pressure” and “Foot Acceleration” do not equally respond according to the three types of audio frequency that are High, Low and Control (Wiederstein and Sippl 2007).
The research indicates that various types of replicated measures and its analytical interpretation are crucial for decision-making. The 95% confidence intervals created the upper and lower bound of the averages and proportions of variables. No statistical inferential test such as ANOVA is applied here. Only by visual observation of the statistic (average or proportion) and its confidence intervals the variability and grouped wise differences are measured. In future analysis, the researcher would like to use statistical inferences and tests for better understanding.
Research Limitations:
All the experimental variables are not included in the data set. Hence, lack of understanding about footwear interface and its effect on preferability may be created. The sample size of this experiment is not also high. Hence, sampling error may be present in this experimental data (Martin and Sayette 1993).
Reference:
Hand, D.J., 2007. Principles of data mining. Drug safety, 30(7), pp.621-622.
Martin, C.S. and Sayette, M.A., 1993. Experimental design in alcohol administration research: limitations and alternatives in the manipulation of dosage-set. Journal of Studies on Alcohol, 54(6), pp.750-761.
Morey, R.D., 2008. Confidence intervals from normalized data: A correction to Cousineau (2005). reason, 4(2), pp.61-64.
Wiederstein, M. and Sippl, M.J., 2007. ProSA-web: interactive web service for the recognition of errors in three-dimensional structures of proteins. Nucleic acids research, 35(suppl_2), pp.W407-W410.
Order an Essay Now & Get These Features For Free:
Turnitin Report
Formatting
Title Page
Citation
Outline
