Abstract

The purpose of the experiment was to test facial tissue tensile strength for various brands.

An apparatus was constructed by which marbles were repeatedly dropped from a fixed height onto a given brand of dry tissue until the tissue failed (marble passed through the tissue). The experiment was then repeated with dampened tissues.

Analysis of the data included dot plots, side-by-side boxplots, and two-way ANOVAs with interaction plots and residual plots.

Report

An experiment was conducted to investigate the strength, perhaps described as tensile strength, of 7 brands of facial tissues. The investigative Tissue Test Team members were Johnny Henderson, Jonathan Ramey, Judiith Salmon, and Joanne Schweinsberg (hereafter known as J4-T3). The 7 brands selected for the experiment were Kleenex (1), Top Crest (2), CVS (3), Kleenex Ultra (4), WalMart(5), Puffs (6), and Puffs Extra Strength (7).

An apparatus was constructed from which a marble (weighing .01 lb) was dropped with initial velocity zero. It traveled a distance of 14.4 cm before striking the surface of a tissue stretched over an embroidery frame. (See figure 1)

DATA COLLECTION

1. A brand of tissue was picked and a tissue was taken . The tissue was stretched over a five-inch embroidery hoop, which was placed horizontally. A marble was then dropped repeatedly onto the tissue until the tissue failed. Failure was deemed to have occured when the marble passed entirely through the tissue. The number of times the marble was dropped was recorded and the process was repeated 10 times for each brand of tissue.

2. The procedure described above was repeated for tissues that had been dampened before dropping the marble. For dampening, a tissue was stretched over an embroidery hoop. As it was held vertically, the tissue was sprayed from a water bottle held 28.8 cm from the hoop. After being dampened, the tissue was placed horizontally, and the marble was dropped. (See figure 2)

POTENTIAL BIAS

There are three main factors:

1. For each brand of tissue tested, tissues were selected successively from one box only.

2. When the marble was dropped repeatedly onto the same tissue, different spots of the tissue were struck.

3. When the tissues were dampened, moisture was more concentrated for some tissues, more uniform for others.

ONE-WAY ANOVA FOR DRY TISSUES

The hypotheses test (using the transformed data) showed that there was a difference in the dry tissue strength due to the tissue type. (Fstat=25.85, p=0.000). I then used the Scheffe multiple comparison test, because that is the one that I am familiar with.

Table of means:

          1 (2.4)   2 (3.11)  3 (1.99)  4 (4.26)  5 (2.35)  6 (1.81)
2 (3.11)   .71  
3 (1.99)   .41       1.12* 
4 (4.26)  1.86*      1.15*    2.27* 
5 (2.35)   .05        .76      .36      1.91* 
6 (1.81)   .59       1.30*     .18      2.45*      .54  
7 (3.40)  1.00*       .29     1.41*      .86*     1.05*     1.59*

The top row and right-hand column contain: tissue type (mean # of marbles). each other entry in the table is absolute value of the difference of the corresponding means; it., .71=3.11-2.40.

The numbers in the table are then compared to the following number when alpha=0.05:

SQRT[(k-1) f(k-1,n*-1,alpha) (MSE) ((ni-nj)/ni*nj)]= SQRT[(6) 2.25 (.2659) (.2)]=.846

If the number in the table is larger than .846 (Note the asterisks), then there is a significant difference in the strengths of the two tissues.

Observations:

• Brand 4 (Kleenex Ultra) differs in strength from all other brands.
• Brand 7 (Puffs Extra Strength) differs in strength from brands 1, 3, 4, 5, and 6. Note that there is no significant difference in strength between brands 2 (Top Crest) and 7. Also, the significant difference between 4 and 7 is not as overwhelming as the others as it disappears if alpha=0.01.
• Brand 2 differs in strength from brands 3 and 6 as well as 4.

Once you know the significant difference exits, the box plots can tell you which is stronger, or a two-sample t-test can be used to investigate any particular pair.

ONE-WAY ANOVA FOR WET TISSUES

The hypotheses test (using the transformed data) showed that there was a difference in the wet tissue strength due to the tissue type. (Fstat=27.857, p=0.000).

Scheffe's multiple comparison test...

          1 (1.39)   2 (1.46)  3 (1.56)  4 (2.12)  5 (1.34)  6 (1.48)
2 (1.46)   .07  
3 (1.56)   .17        .10 
4 (2.12)   .73*       .66*      .56* 
5 (1.34)   .05        .12       .22       .78* 
6 (1.48)   .09        .02       .08       .64*      .14  
7 (2.56)  1.17*      1.10*     1.00*      .44      1.22*     1.08*

SQRT[(6) 2.25 (.0753) (.2)]=.451

Observations:

• Both brand 4 and brand 7 are significantly stronger than brands 1, 2, 3, 5, and 6. However, there is no significant difference between brands 4 and 7.

CONCLUSIONS

If you want a tissue that will not fail you under both wet and dry conditions, then go with brands 4 or 7, admittedly the more expensive of the tissues we tested. If you are only concerned about dry conditions and price might be a concern then brand 2 might be good enough...just don't sneeze into it unless you want a wet hand!

POSSIBLE EXTENSIONS

1. Re-do the analysis without the most extreme cases --- Kleenex Ultra and Puffs Extra Strength.

2. Do more trials with more brands.

3. Develop a method for spraying that would more effectively control the concentration of water.

4. Investigate the percent of each brand that failed at a specified height.

5. Add additional variables and extend the ANOVA analysis.

Comment: Testing for softness rather than strength would probably be more meaningful to the consumer.

Return to Top

Link to Accompanying Dataset

Return to STATS Project Reports Index