RESULTS: SURVEY

There were 79 complete responses to the phone survey. An additional 10 households contacted by phone refused to participate and 2 gave incomplete responses. Thirty-three of the 80 households receiving the mail survey gave responses. However, due to a clerical error, 17 of the mail responses received could not be identified with a home or toilet. A follow-up letter convinced four of the mail respondents to re-submit a survey, however 13 did not respond. This leaves 20 usable responses from the mail survey, for a total of 99 usable responses from both phone and mail.

Seventy-eight percent of the survey respondents were the occupants at the time that the rebate toilets were purchased for the house. Based on phone book listings and tax records it appears that 66 percent of all possible respondents were the occupants at the time the rebate toilets were purchased.

A comparison of participation rates for original occupants versus new owners confirms that homes with respondents who were the original purchasers of the low-consumption rebate toilets responded to the survey at a greater rate than homes with new owners since the rebate program expired. Of the possible respondents identified as not being the occupants at the time the toilets were installed, 39 percent responded to the survey. Sixty-eight percent of original occupants gave responses. This difference is large and statistically significant.

Possible Sources of Bias

The finding that original occupants responded at a greater rate than those who moved in later is not surprising, since those who participated in the toilet rebate program are familiar with questions about toilets purchased under the rebate program, while those who were not the original occupants may know nothing of the rebate program or even which toilets in their house are 1.6 gallon models. Also, phone numbers for original occupants were given in the data from Tucson Water, while phone numbers for those who were not original occupants were available only if they were listed in the phone book.

Survey results from original occupants may be susceptible to bias in favor of water conservation. All of the homes selected had owners who willingly participated in a water conservation program. To the extent that they were motivated to save water over the need to replace a toilet they would have needed to replace anyway, these participants may be said to be biased towards viewing water conservation programs and water saving devices favorably. They may ignore minor inconveniences in the functioning of their low-consumption toilets because they are more influenced by the thought that they are saving water. Comments provided by respondents confirm that this attitude is true for at least a couple of participants.

However, many of the homeowners from the time of the rebate program have since moved out of the homes. The new occupants are less likely to be biased towards viewing water conservation favorably. Also, earlier investigations into the toilet rebate program revealed that at least some of the program participants were motivated by the fact that they needed to replace their toilets, and would have done so anyway without the toilet rebate program. These two factors serve to balance against the possible bias in favor of water conservation.

Survey Responses

On average, respondents had 1.89 adults and 0.53 children occupying their households on a regular basis. The average persons per household of 2.42 is only slightly lower than the average persons per household reported in the 1990 census for the Tucson Metropolitan Area and is consistent with the trend of decreasing household size.

The average rating for all respondents of low-consumption toilets purchased through the Tucson Water toilet rebate program was 3.83 on a five point scale (131 toilets). This shows that respondents on average were satisfied with the functioning of their rebate program toilets. By comparison, the overall rating for toilets in the Los Angeles Department of Water and Power Survey in 1992 was 7.4 on a 10 point scale (Wirthlin Group 1992). If this rating is halved to correspond to a five-point scale, the overall rating was 3.7, or slightly lower than in the current survey.

Respondents overall rated their non-rebate toilets slightly higher than the rebate toilets at 3.92 (53). Of the non-rebate toilets, respondents rated their non-rebate, non-low-consumption toilets (those using 3.5 gallons per flush or greater) (3.93, 44 toilets) slightly higher than their non- rebate low-consumption toilets (3.88, 8 toilets).

Figure-3. 95 Percent Confidence Interval Around Mean Toilet Ratings

There was only one toilet for which more than 30 responses were received. The average rating of the Wellworth Lite was 3.82 (51 toilets). Other models with ratings for more than 5 toilets are the Atlas at 3.87 (15 toilets), the Aqua Saver at 4.44 (9 toilets), the Quantum (5.0, 7 toilets), the Allegro (3.43, 7 toilets), the Lamosa Sahara (3.57, 7 toilets), and the Preserver (3.67, 6 toilets). These ratings are shown with 95 percent confidence intervals around the means in Figure 3. Only the rating for the Aqua Saver and the Quantum can be said to be greater than the mean for all rebate toilets at the 95 percent confidence level. However, it must be kept in mind that the ratings for these toilets are based on small numbers. Also, none of the ratings for each toilet are statistically different from each other, except for the Quantum which can be said to be significantly better rated than all other brands for which there were at least 5 toilets except the Aqua Saver. The sample size for the survey would need to be significantly increased for toilets other than the Wellworth Lite in order to rank toilet ratings in a statistically significant manner.

There were four brands of pressurized toilets for which data was collected in the survey - the Quantum, the Rialto Lite, the Cadet, and the Ultra-Flush. The survey yielded ratings for 14 pressurized toilets. Every pressurized toilet was given a rating of 5, the highest rating. However, the survey also revealed that 4 Quantum pressurized toilets in 3 households were removed due to bad performance or lack of available replacement parts. By contrast, only 2 other toilets in 1 household were reported to be removed. They were yielded inoperable by actions of children in the household. It appears from this limited sample that when pressurized toilets work well, residents are very happy with them, but when they do not perform correctly, residents are sufficiently dissatisfied to remove them.

Respondents were asked specifically about their flush mechanisms. Of those respondents who knew whether or not their toilet had a flapper as the flush valve, 58 percent had replaced the flapper on their rebate toilets. Eighty-five percent of those replacing their flappers had gone to a hardware store for a replacement flapper. Nine percent got replacement flappers from their plumbers, while six percent got replacement flappers from a specialty toilet supply store.

Respondents were asked whether or not they use in-tank bowl cleaners. However, no attempt was made to distinguish between halogenating cleaners and non-halogenating cleaners. Twenty- four percent of those who knew they had flappers as toilet flush valves used in-tank bowl cleaners. Several respondents stated they could notice the deteriorating effect of using in-tank bowl cleaners on their flappers and had stopped using this type of cleaner for that reason, or that plumbers or toilet salespeople had warned them not to use in-tank bowl cleaners.

Reported Problems

In the survey, respondents were asked whether there were chronic or recurring problems with any of their toilets. Four possible problems related to toilet functioning were suggested: frequent double-flushing, clogging, toilet flushes or refills on its own, or toilet runs after flush and then stops. The most numerous problem reported with both rebated low-consumption toilets and non- low-consumption, non-rebated toilets was double flushing. Multiple flushing was reported as necessary in 38 percent of rebate toilets and 16 percent of non-rebate, non-low-consumption toilets. This difference is statistically significant. Clogging was reported in 28 percent of rebate toilets and 8 percent of non-rebate, non-low-consumption toilets. This difference is also statistically significant. A toilet flushing or refilling on its own was reported in 16 percent of rebate toilets and 18 percent of non-rebate, non-low-consumption toilets, which is not a statistically significant difference. Running after the flush and then stopping was reported in 25 percent of rebate toilets and 16 percent of non-rebate, non-low-consumption toilets. This is not a statistically significant difference.

Reports of double flushing in a rebate toilet match the data logging record in 67 percent of the cases. This percentage stays the same whether the double flush percentage used in the data logging is 15 or 10 percent of flush events. The percent of toilets requiring double flushing as reported in the survey (38%) is higher than the percentage requiring double flushing obtained from the data logging (10%).

Table 3. Percent of Toilets Reported to Have Chronic Toilet Function Problems